Kitua, Gabriel Daudi Yona2019-10-032020-01-082019-10-032020-01-081996Kitua, G. D. Y. (1996). Application of multiple discriminant analysis in developing a commercial Banks loan classification model and assessment of significance of contributing variables: a case of National Bank of Commerce. Master dissertation, University of Dar es Salaam. Available at (http://41.86.178.3/internetserver3.1.2/search.aspx?formtype=advanced)http://localhost:8080/xmlui/handle/123456789/4717Available in print formThe purpose of this study was to develop Commercial Bank's loan classification model and asses the relationship between relevant variables and existing bad loan problem. Multiple discriminant analysis was applied. The analysis was based on March 31st, 1994 overdraft loan accounts data. The classification model developed shows that there exists a linear relation between loan classification and the six variables considered. Three functions below were developed. S1=-4 . 09 +. 0001x1-. 0365x2+: 18x3+.2407x4+.017 8x5+5. 329x6 S2=-4.4335+.0007x1+.0157x2+.233x3+.253x4+1.007x5+.195x6 S3 =-5.255-.0009X1+.018x2+1.4085x3-.149X+.104X5+.556XX6 Where: S;: Loan score as estimated by function i, xl: Number of days the loan is past due, x2: Number of years the Borrower have been in the same business, x3: Borrower's cooperation and management capability, x4: Collateral, x5: Operation of borrower's account and x6: Propriety of use of borrowed funds. Groups' scores cut-off points with respect to the three functions above are shown in the table below: LOSS l DOUBTFUL 1 SUBSTANDARD + ESP.MENTIONED 1 CURRENT -2.4569 -.2846 1.8921 2.13478 -1.607 -.8705 .0658 1.4099 -1.0424 .Oll .7362 1.2606 Thus S3, for example, will signify a substandard classification if it lies between .011 and .7362. Bayesian decision making approach is applied in determining the most likely score out of three scores from the three functions. Propriety of use of funds borrowed; operation of Borrower's overdraft account; cooperation with the Bank and Borrowers management capability are important factors in determination of the quality of loan portfolio. Other less important variables are collateral and number of days the loan is past due. A variable discriminating between old and new borrowers was least important but indicates that borrowers with experience in business financed by the Bank are less risk than inexperienced customers. The model can be used by NBC loan officers in the loan review process and by Bank Examiners in loan classification of existing loan portfolio. Groups' scores cut-off points with respect to the three functions above are shown in the table below: LOSS l DOUBTFUL l SUBSTANDARD ! ESP.MENTIONED j CURRENT Sl : -2.4569 -.2846 1.8921 2.13478 S2 : -1.607 -.8705 .0658 1.4099 S3 : -1.0424 .011 .7362 1.2606 Thus S3, for example, will signify a substandard classification if it lies between .011 and .7362. Bayesian decision making approach is applied in determining the most likely score out of three scores from the three functions. Propriety of use of funds borrowed; operation of Borrower's overdraft account; cooperation with the Bank and Borrowers management capability are important factors in determination of the quality of loan portfolio. Other less important variables are collateral and number of days the loan is past due. A variable discriminating between old and new Borrowers was least important but indicates that Borrowers with experience in business financed by the Bank are less risk than inexperienced customers. The model can be used by NBC loan officers in the loan review process and by Bank Examiners in loan classification of existing loan portfolio.enBank loansCommercial loansTanzaniaThesis