Comparisons Among Credit Evaluation Methodology for Micro Enterprise : Evidence from Credit Evaluation Model for Micro Enterprise of KOREG in Korea

Abstract

The purpose of this study was to understand which variable can most effectively benchmark subrogation and use the result to establish a more stable guarantee policy. In this study, credit ratings are calculated from the model to evaluate micro-enterprises using data mining technique for micro-enterprises. The model to evaluate micro-enterprises is obtained from the local credit guarantee foundation. With respect to prediction accuracy of each of the models from the study result, out of decision tree, CHAID was 67.7% and it is the highest model prediction accuracy and prediction of subrogated performance using CHAID model out of decision tree can bring good result of prediction. In addition, because in CHAID model of decision tree the combination grade for which added points are not applied was 79% and this shows predictor importance, it is necessary that the combination grade for which added points are not applied predicts subrogated performance as the optimum predictor and uses the result of prediction for guarantee policy. Suggestions through the study result are like the following. The result of prediction of occurrence of subrogated performance for guaranteed enterprise through evaluation model for Micro enterprisers should be reflected to guarantee supply policy to lower the rate of subrogated performance and through these basic materials that make the tax of government used efficiently are provided. Because of this, suggested is the appropriateness that each of the local foundations can predict subrogated performance through credit evaluation model and use the result of prediction for adjustment of guarantee rate before supplying guarantee, and Korean Federation of Credit Guarantee Foundations should decide whether the model to evaluate credit of Micro enterprisers was reflected or not in audit of supplementation through reguarantee. In addition, likewise Credit Scoring System(CSS) used for individual guarantee should reflect the result of prediction of subrogated performance and decide guarantee limit or whether or not of approval of guarantee. Therefore, the outcome of this study would have great significance as the method to reduce guarantee risk.