|摘要: ||本研究主要目的為探討農會信用部授信評估之影響因素，並以銀行授信風險評估5P原則為基礎，即借款人、借款用途、還款來源、債權保障與借戶展望等因素，來評估授信戶的信用與風險，以羅吉斯迴歸模型(Logistic Regression Analysis)做為本研究的統計分析方法。|
The main purpose of this study is to explore the factors influencing credit evaluations in the credit department of Farmers’ Association; and base on the 5P principle of bank’s credit risk evaluation, such as people, purpose, payment, protection, and perspective, the borrower’s creditability and risk are evaluated. The Logistic Regression Analysis is adopted as the statistical analysis method for this study.
The scope of the research data includes land loans, housing loans, and loans of strategic agricultural projects by one of the Farmers’ Association in Changhwa County. With loans between January, 2010 and December, 2012, excluding bad credits and disqualifications, there are 51 loans which overdue for more than three months and are included in the range of overdue loans; the remaining 422 loans belong to normal repayments. Total sample size is 473 and the Logistic Regression is used for data analysis. There are sixteen variables in the study and they are respectively,membership, gender, age, education, marital status, occupation, credit status, loan status, loan purpose, annual income, collateral ownership, guarantor, property status, loan amount, loan-to-value, and interest rate.
The empirical results show that the overall accuracy of the model is 97.9%; the variables significantly influence the credit evaluation factors, under the significance level of 10%, there are eight variables that show significant correlations. Among them, the “loan status”, “loan amount”, and “interest rate” are significantly and positively correlated with the probability of overdue payment. The five variables of “age”, “annual income”, “property ownership”, “guarantor”, and “property status” are significantly and negatively correlated with the probability of overdue payment. The eight less significant variables are “embership”, “gender”, “education”, “marital status”, “occupation”, “credit status”, “loan purpose”, and “loan-to-value”.
The empirical model of this study can provide Farmers’ Association with references for credit evaluations and can effectively control credit quality, reduce credit risks and losses.