Factors Influenced on the Bank Performance - Research Paper Example

Assignment The performance of a bank is influenced by many factors which include; core deposits, loans to organizations and individuals, operating income and revenues, non-performing assets, securities, volatile liabilities and employee expenses. In addition, declining provisions which is way below the charge-offs leads to improved credit quality while net interest continues to increase. The influence is further compounded by several risks facing them which include slowdowns in real estate lending and common market investments as well as slow growth deposits growth. In this research study, the impact of total assets, income earning assets, deposits, equities and expenses on the total bank revenue was to be determined. Each variable’s impact was defined holding all the others constant. Then regression equations of the nature; R = C + β0Es and R = C + β0I were formulated (where C = a constant, E = Expenses and I investments, β0 = coefficients). Table of Contents Abstract 2 Literature Review 4 Methodology 4 Data 4 Analysis 5 Results and Discussions 5 Conclusions and Recommendations 12 References 12 Appendix-Log File 13 List of Tables Table 1: Banks by Country 4 Table 2: Banks by Specialization 5 Table 3: Coefficients Table 5 Table 4: Coefficients Table 7 Table 5: Coefficients Table 8 Table 6: Coefficients Table 10 Table 7: Means of Revenue Vs Specialization 11 List of Figures Figure 1: Residual Plot 6 Figure 2: Fitted Values 8 Figure 3: Residuals Fitted 9 Figure 4: Best Line of Fit 10 Literature Review A banks performance according to Federal Financial Institutions Examination Council (FFIEC, 2004) is influenced by many factors which include; core deposits, loans to organizations and individuals, operating income and revenues, non-performing assets, securities, volatile liabilities and employee expenses. To FFIEC, this manifestation is see through the way the large banks perform as opposed to the small banks since the large ones have many performing assets and also since their loan provisions are low as compared to the small banks with low client base and trust. Declining provisions which is way below the charge-offs leads to improved credit quality while net interest continues to increase. On the other hand, banks face several risks which include slowdowns in real estate lending and common market investments as well as slow growth deposits growth. Further, there is the possibility of full arrests in their operations due to structured long term (Altunbas and Ibáñez 2004). According to Pennathur, A. at el. (2001), non-interest income in a bank is strongly influenced by its size, who owns it and the management in place. A bank might be big enough but what the returns translate into matters much but studies done on commercial banks show that non-interest income forms a large part of all banks operating profits. Further, according to Barnes (2004), the growth in mutual deposits at banks is slowly going down since customers have other alternatives, to save their money which include mutual investments and cash accounts, which offer transactions/services similar to those offered by banks and which also pay better. Operations cost (total expenses) is another factor which determines which banks get as the net income at the end of each financial period (Richmond, 2006). According to Richmond, this is so because, the higher the operating cost the lower the savings. However, a bank may have higher operating costs due to expansion plans with the aim of a long term benefits which at times do no translate into better earnings in the long run (Hill 1992). Methodology In this research study, the impact of total assets, income earning assets, deposits, equities and expenses on the total bank revenue was to be determined. Each variable’s impact was defined holding all the others constant. Data The population size is N = 86,460 and a representative sample was calculated using the following formula: n = N / (1+N (e2)). The expected error is at least 5%; n = 86,460 / (1+86,460* (0.052)) therefore n = 398.17 and the sample size is approximately 400. Since all the banks are not the same, then a stratified sample is selected which represent all the banks depending on the specialization and country code according to the table 1 and 2 below. This was dictated by the arguments of Freed, M. et al. (1991) on sample selection and population representation. Table 1: Banks by Country Country Total Proportion Sample AT 5,340 0.061763 25 BE 2,520 0.029146 12 DE 40,590 0.469466 186 ES 4245 0.049098 20 FI 495 0.005725 2 FR 10,485 0.12127 49 GR 780 0.009022 4 IE 1245 0.0144 6 IT 14,685 0.169847 68 LU 2,670 0.030881 12 NL 2,085 0.024115 10 PT 1,320 0.015267 6 Total 86,460 1 400 Table 2: Banks by Specialization Specialization Total Proportion Sample Bank Holding & Holding Companies 2,085 0.024115 11 Commercial Banks 26,175 0.302741 121 Cooperative Bank 39,420 0.455933 181 Investment Banks 1,455 0.016829 7 Real Estate / Mortgage Bank 2,805 0.032443 13 Savings Bank 14,520 0.167939 67 Grand Total 86,460 1 400 Analysis Data analysis was done using Stata and descriptive statistics, regression analysis and graphs were used to interpret the data. This was supported by Hall and Mishkin argument that the best way to investigate the contribution of certain factors is the use of regression analysis (both simple and multiple). The results follow below; Results and Discussions Table 3: Coefficients Table From the table above, the regression equation can be formulated as below; R = -33.98 - .32A + .46 L + .41 S - .06D + .49 E…………………………..1 (Where R = Revenue, A = Assets, L = Loans, S = Securities, D = Deposits and E = Equity) This means that a unit increase in assets adjusting for the other variables leads to .32 units decrease in revenues. A unit increase in Loans adjusting for the other variables leads to .46 units increase in revenues. Further, a unit increase in securities adjusting for the other variables leads to .41 units increase in revenues. A unit increase in deposits adjusting for the other variables leads to .06 units decrease in revenue while a unit increase in equity adjusting for the other variables leads to .49 units increase in revenue. Predicting the residuals for the model and plotting them as shown below show that they are not normally distributed (see figure 1). Figure 1: Residual Plot Heteroskedasticity Test Breusch-Pagan / Cook-Weisberg test for heteroskedasticity Ho: Constant variance Variables: fitted values of revenue chi2 (1) = 1404.82 Prob > chi2 = 0.0000 The test shows that the chi-square value is big (1404.82) indicating heteroskedasticity is a problem. The variables revenue, assets, loans, securities, deposits and equity are significant at 95% level of significance. Considering total revenues as the dependent variable and total assets as the independent variable then; Table 4: Coefficients Table The regression equation is as shown below; R = 49.98 + .06A……………………..2 This means that a unit increase in assets leads a .06 unit’s increase in revenue. The fitted line is as shown below; Figure 2: Fitted Values Heteroskedasticity Test Breusch-Pagan / Cook-Weisberg test for heteroskedasticity Ho: Constant variance Variables: fitted values of revenue chi2 (1) = 1091.09 Prob > chi2 = 0.0000 The test shows that the chi-square value is big (1091.09) indicating heteroskedasticity is a problem. The variables revenue, assets, loans, securities, deposits and equity are significant at 95% level of significance. Considering revenue together with interest expense, personnel expenses and loan loss provisions and carrying out a multiple regression yields the following coefficients; Table 5: Coefficients Table R = 8.33 + 1.04 IE + 2.15 PE + .16LP……………………………………..3 (Where R = Revenue, IE = Interest Expenses, PE = Personnel Expenses and LP = Loan Provision) From the equation, a unit increase in interest expenses leads to 1.04 units increase in revenue. A unit increase in personnel expenses leads to 2.15 units increase in revenues while a unit increase in loan provisions leads to .16 units increase in revenue. Figure 3: Residuals Fitted The residuals fit shows that the residuals are not normal and thus goes against the assumptions of normality thus normal non-parametric tests are performed. Heteroskedasticity Test Breusch-Pagan / Cook-Weisberg test for heteroskedasticity Ho: Constant variance Variables: fitted values of revenue chi2 (1) = 1749.02 Prob > chi2 = 0.0000 The test shows that the chi-square value is big (1749.02) indicating heteroskedasticity is a problem. The variables revenue, assets, loans, securities, deposits and equity are significant at 95% level of significance. Considering the total expenses and revenue alone yields the following; Table 6: Coefficients Table R = -6.17 + 1.09E…………………………….4 A unit increase in expenses leads to a unit increase in revenue. Figure 4: Best Line of Fit According to Gore and Altman (1992), multiple comparisons are used to show the difference in means for different groups. A non-parametric test of mean differences in the banks according to the country codes shows significance difference [Chi-square = 51.02, p = .0000] (see below) A non-parametric test of mean differences in the banks according to the specialization shows significance difference [Chi-square = 39.17, p = .0000] (see below) Table 7: Means of Revenue Vs Specialization Conclusions and Recommendations From the above analysis, it is evident that at times higher expenses mean better revenues since they may be as a result of expansion programmes. Further, heteroskedasticity test showed that the chi-square values are big indicating heteroskedasticity is a problem and thus normality tests could not be performed. In order for a bank to have a niche, the following are recommended originating from the analyzed results; Loan provisions should be made low as possible; The higher the assets, the better the net income thus banks should strive to maintain a high assets portfolio. References Altunbas, Y and Ibáñez, M. (2004). Mergers and Acquisitions and Bank Performance in Europe. The Role of Strategic Similarities. Working Paper Series, No. 398, October Barnes (2004). Loans and Equities in the Performance of Banks. Nairobi: General Printers Ltd Federal Financial Institutions Examination Council (FFIEC). Condition and Performance of Commercial Banks. Quarterly Journal, Vol. 23, No. 1, March 2004 From panel data of households”, Econometrica, 50, 261-81. Freed, M. et al. (1991). Handbook of Statistical Procedures and Their Computer Applications to Education and the Behavioural Sciences. NY: American Council on Education, Macmillan Publishing Company Gore, S.M. & Altman, D.G. (1992). Statistics in Practice. London: British Medical Association New York: Wiley Hall, R., and F. Mishkin, (1982), “The sensitivity of consumption to transitory income: Estimates Haron, S. (2005). The effects of management policy on the performance of Islamic banks. Springer Netherlands, Asia Pacific Journal of Management, 63-76 Hill, M. (1992), The Panel Study of Income Dynamics: A user’s guide, Newbury Park, California: Sage Publications. Pennathur, A. at el. (2001). Does non-interest income impact bank performance in emerging markets? The case of India. [Online] Available from: http://www.fma.org/Texas/Papers/noninterestincomeandbankperformanceinindia.pdf Richmond (2006). Commercial Banks: The Risks Involved. John Hopkins University. Baltimore Appendix-Log File --------------------------------------------------- log: C:\Documents and Settings\HP LTop\My Documents\freelance work\econom > etrics.smcl log type: smcl opened on: 1 Feb 2010, 10:06:18 . reg revenue assets loans securities deposits equity, beta Source | SS df MS Number of obs = 400 Model | 1.3185e+09 5 263701351 Prob > F = 0.0000 Residual | 31712983.6 394 80489.8062 R-squared = 0.9765 -------------+------------------------------ Adj R-squared = 0.9762 Total | 1.3502e+09 399 3384009.37 Root MSE = 283.71 revenue | Coef. Std. Err. t P>|t| Beta assets | -.3185681 .0404268 -7.88 0.000 -4.95129 loans | .4683312 .0429002 10.92 0.000 2.778005 securities | .4152397 .0403834 10.28 0.000 3.757537 deposits | -.0589454 .0095694 -6.16 0.000 -.7466671 equity | .494335 .0538412 9.18 0.000 .2504983 _cons | -33.98454 15.7508 -2.16 0.032 . . predict res, resid . histogram res, norm bin(50) (bin=50, start=-1853.3842, width=66.728125) . line assets revenue . mspline assets revenue unrecognized command: mspline . line assets revenue . twoway (scatter revenue assets)||fpfitci revenue assets . twoway (scatter revenue assets)||fpfitci revenue assets, xtitle("Assets") ytitl > e("Revenue") note("European Banks") . hettest Breusch-Pagan / Cook-Weisberg test for heteroskedasticity Ho: Constant variance Variables: fitted values of revenue chi2(1) = 1404.82 Prob > chi2 = 0.0000 . test assets ( 1) assets = 0 F( 1, 394) = 62.10 Prob > F = 0.0000 . reg revenue assets Source | SS df MS Number of obs = 400 Model | 1.2712e+09 1 1.2712e+09 Prob > F = 0.0000 Residual | 79046440.3 398 198609.146 R-squared = 0.9415 -------------+------------------------------ Adj R-squared = 0.9413 Total | 1.3502e+09 399 3384009.37 Root MSE = 445.66 revenue | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- assets | .0624287 .0007803 80.00 0.000 .0608946 .0639628 _cons | 49.98296 23.77838 2.10 0.036 3.236042 96.72988 . reg revenue assets, beta Source | SS df MS Number of obs = 400 #NAME? Model | 1.2712e+09 1 1.2712e+09 Prob > F = 0.0000 Residual | 79046440.3 398 198609.146 R-squared = 0.9415 -------------+------------------------------ Adj R-squared = 0.9413 Total | 1.3502e+09 399 3384009.37 Root MSE = 445.66 revenue | Coef. Std. Err. t P>|t| Beta assets | .0624287 .0007803 80.00 0.000 .9702869 _cons | 49.98296 23.77838 2.10 0.036 . . hettest Breusch-Pagan / Cook-Weisberg test for heteroskedasticity Ho: Constant variance Variables: fitted values of revenue chi2(1) = 1091.09 Prob > chi2 = 0.0000 . kwallis revenue, by(cc) variable cc not found in option by() - label var cc "Country Code" - label var spec "Specialization" . kwallis revenue, by(cc) . tab cc, sum(revenue) . kwallis revenue, by(spec) . reg revenue intexpense perexpense llp . reg revenue intexpense perexpense llp, beta . predict res1, resid . histogram res1, norm bin(50) . hettest . reg revenue expenses . reg revenue expense . reg revenue expense, beta . twoway (scatter revenue expense)||fpfitci revenue expense, xtitle("Expenses") y > title("Revenue") note("European Banks") Read More