Evaluation And Prediction Of Township Government Fiscal Risk

The operation of Country economic development was carried out by the township govermentfiance, which plays the key role in the social and economic management. Recently, the townshipgoverment has faced serious problems, such as heavy debts, increasing deficit and high riscal risks.Therefore, it is very important to evaluate and predict the fiscal risk of township goverments.This thesis started from the fuction and present status of township goverment fiance,summarized the characteristics of the fiscal and debts risks of Chinese township governments, andanalyzed the principle reasons from the views of economic system, fiscal system and debtsmanagement.Using the reference of early experiences for finicial risk prediction, this thesis established theintegrated evaluating index for the township government fiscal risks, which are coupled by the relatedand respectively isolated single indexes, including: fiscal deficit ratio, debts burden rate, debts ratio,debts dependence, debts growth rate, debts service rate and debts overdure rate. The integratedevaluating indexes are comprosed by the single indexes within their own weights, which are gainedfrom the application of Analytical Hierarchy Process Method.The prediction model based on the BP neural network method is established and compiled usingMatlab langurage, which contains the Graphic User’s Interface. The model has been verified by thesample data analysis, computing and valiation. After the modification of every parameter, the desiredindex will be predicted.In order to demonstrate the evaluating and predicting model here, eight township governmentsare selected as the samples. The weights of every single evaluatiing index have been gained using theAHP method. The sampling data from2005to2010are employed as the input, the data of2011areused as the validating basis to train the BP neural network predicting model. Finially, the indexes of2012are calculated and proposed as the accordance of the prevention means.Based on the calculation and analysis of the present evaluating and predicting model, it can be seen that the integrated evaluating index can avoid the randomness and usteadiness of single indexes,which can represent the fiscal status of the township government and the risk grade of fiscal and debts.The BP neural network prediction model has shown its reliability and practicability and also has moreroom to be improved.