Research On Target Price Insurance Pricing And Optimization Of Yunnan Pu ’er Coffee Beans

As one of Yunnan’s special agriculture,the coffee industry plays an important role in the national coffee production.In recent years,due to the influence of international coffee trade giants,coffee prices fluctuate significantly,and the interests of coffee farmers and enterprises are not protected,which seriously restricts the sustainable and healthy development of Yunnan coffee industry.Therefore,in 2016,the People’s Government of Pu ’er City issued the Pilot Program of Policy-based Coffee Price Insurance in Pu ’er City,and took the lead in launching the first coffee price insurance pilot in China in Ning ’er County,aiming at stabilizing the coffee price and guaranteeing the profits of coffee farmers and enterprises.However,according to the pilot situation,there are many problems in coffee price insurance,such as the insurance period,target price,insurance amount and rate are not reasonable.Among them,the most prominent is that the target price is not reasonable,which further leads to the deviation in the insurance rate determination,and directly affects the interests of farmers and insurance companies.Therefore,it is necessary to explore reasonable target price and further optimize coffee price insurance,protect the interests of policyholders and insurers,and promote the high-quality development of coffee price insurance.In order to solve the above problems,on the basis of analyzing the price risk of Yunnan coffee beans,this paper uses ARIMA model,ARIMAX model,single feature LSTM neural network,multi-feature LSTM neural network and other algorithms to predict the coffee price,determine the target price,and determine the rate,and optimize the design of coffee target price insurance products.The research contents of this paper are as follows: First,the Census X12 seasonal adjustment method and HP filter method are used to decompose the coffee market price in Yunnan Province,and the risk characteristics of the price fluctuation of high-quality coffee in Yunnan province are analyzed.The results show that the price of high-quality coffee in Yunnan is affected by seasonality,periodicity,randomness and tendency.Secondly,based on the pilot situation of coffee target price insurance,the appropriate insurance period is discussed.At the same time,aiming at the problem of unreasonable target price setting,various algorithms such as ARIMA model,ARIMAX model,single feature LSTM neural network and multi-feature LSTM neural network are used to predict the coffee price.Based on the root mean square error(RMSE),the prediction accuracy of various algorithms was evaluated,the optimal prediction model was selected,and the appropriate target price was finally determined.Finally,based on the current implementation plan of target price insurance,the expected loss method is adopted to determine the premium rate of target price insurance,optimize the insurance amount design,and determine the premium rate.The empirical study shows that the multi-feature LSTM neural network has the best prediction effect,and the target price obtained in this paper is closer to the actual price set by the original scheme.At the same time,the multi-feature LSTM neural network can make full use of various variables related to coffee transaction price,and finally achieve accurate coffee target price insurance rate determination.This thesis consists of five parts: the first chapter is an introduction,which mainly describes the background of the subject,the summary of literature,research significance,content and methods,etc.The second chapter mainly introduces related theories and models;The third chapter mainly introduces the situation of green coffee bean industry,price risk and the pilot situation of coffee target price insurance in Yunnan Province.The fourth chapter mainly analyzes the price risk characteristics of Yunnan high quality coffee beans and the pricing research based on the optimal model.The fifth chapter is the conclusion and suggestion of this paper.