Identify the factors influencing on granting credit in lending institutions

Abstract

Credit scoring system based on various criteria is widely used analysis technique of measuring customer credit worthiness by the Credit Information Bureaus (CIB) in many developed countries. However, such a system is not available in Sri Lanka. The aim ofthis study is therefore to identify the factors that influencing on granting credit in lending institutions in Sri Lanka and to recommend a criteria to identify potential boiTowers. The required variables for the study were identified from the Bureau point of view and secondary data were collected from customer credit reports. Further, 75 customers who obtained credit were selected randomly and based on their credit profile eight important variables were derived. The combination of Principal Component Analysis (PCA) and Factor Analysis (FA) was carried out to develop factors. The suitability of data for PCA was confirmed using Bartlett’s Test (X2= 206.092, p= 0.000) and suitability for FA was confirmed by KMO statistic (KMO=0.532). The results of FA showed that the initial system of 8 variables can be reduced to 2-D system which account for nearly 52% of the initial variability. Ofthe identified two principal components, first component came from the, “own credit history” and the second was came from "guaranteed credit history". In order to have better interpretation the results obtained from PCA were rotated using orthogonal factors namely Varimax, Equimax and Quatimax and found that the results obtained were invariant on the type of rotation. Thus it can be concluded that, own credit history and the guaranteed credit history are the influential factors on credit granting. Depending on the variance representing by each factor, it can be concluded that the own credit culture component is more significantly influence on the credit evaluation process than the guarantee component. However, in order to have the accurate credit decision on granting credit, it is required to consider both factors together by indentifying the importance depending on the representing variance. The finding in this study could be effectively employed to develop a suitable credit scoring system for the lending institutions. The study further showed the benefit of using multivariate techniques to solve some problems in lending institutes.