Research On The Improvement Of Farmer Credit Rating System In Gansu Rural Credit Union

In the past 40 years of reform and opening up,China's rural funds have been drawn away.Today,financial institutions act as the main force of drawing rural funds.There are many reasons for the negative flow of rural funds,the most important reason is that the risk of rural household loans is large,and the cost of issuing loans in rural areas is very high.On one hand,the problem of rural land property rights makes the farmers unable to provide sufficient collateral.On the other hand,the complexity of the credit evaluation of farmers makes it difficult for banks to effectively identify the credit risks of farmers.Therefore,it is crucial to establish a set of locally applicable and accurate forecasting household credit rating system.Gansu Rural Credit Union has been serving the three rural areas.It has established a credit rating system for farmers many years ago.This system has been on the line for many years and has shown certain deficiencies in the rating methods,indicator systems and rating procedures.It is imperative to update and improve it.This paper uses theoretical analysis combined with descriptive statistical analysis methods to first explain the three aspects of the current rural household credit rating system in Gansu Province.Then,the farmers who have issued loans from Gansu Rural Credit Union is selected as research objects.The data of 10,000 households are selected as research samples,and the current rural credit Union's rating index and previous research results of Gansu Province are selected to select individual and family characteristics and solvency.39 indicators in the six aspects of credit status,management ability,cooperation status and macro environment are independent variables.Logistic regression model is used to analyze the data of 9500 households in the training sample,and the model of the remaining 500 households is used for model testing.The results show that there are 16 indicators that have significant impact on farmers' default probability.The accuracy rate of the new model for training samples is 89.3%,and the accuracy rate for the test samples is 82.2%.The recognition accuracy rate for good customers is 82.8%.The recognition accuracy rate was 76.9%.It can be seen that the model has a good predictive effect and has certain practical value.Finally,from the perspective of practical operation,this paper proposes four improvement strategies for the credit rating system of rural credit Union in Gansu Province.The research in this paper clearly points out that the factors such as dependent population,external guarantee,neighborhood relationship,personal preference,and cooperative relationship have no significant impact on the credit status of farmers.At the same time,the model is supplemented with marital status,education status,political appearance,number of family labor,and fixed.Significant indicators of the impact of assets on farmers' credit.The use of the new probabilistic forecasting model can more accurately evaluate the credit status of farmers,and provide a more powerful basis for credit officers to make decisions on farmers' loans.This is of great significance to the control of the credit risk of rural households in rural credit Union in Gansu Province,the difficulty in solving the problem of rural household loans in Gansu Province,and even the control of the outflow of rural capital in Gansu.