Empirical Analysis Of Domestic Carbon Financial Transaction Risk Based On Financial Time Series Analysis

Human will inevitably generate a large amount of greenhouse gases in production and life,and the temperature increase caused by excessive emissions will lead to multiple problems,and the negative impact on the climate environment,biodiversity and human society can no longer be ignored.The establishment of carbon emissions trading system is an international means to deal with climate change,and it is the way to minimize the overall emission reduction cost of society.my country began to participate in the carbon financial market through the International Clean Development Mechanism(CDM)project in 2005;since 2013,the pilot carbon markets have been operating one after another;In 2021,the national carbon emissions trading market will be put into use for the first time.The risks of carbon finance mainly include political risk,policy risk,market risk and investment risk.However,all kinds of risks will eventually be reflected in the transaction price.By studying the changes in the transaction price,we can obtain the current operation of the financial market and summarize the possible deficiencies in the current market.At present,the relevant research results at home and abroad mainly focus on the mature European carbon emission trading market,and the relevant research on the domestic carbon market is still in the initial stage.In terms of research methods,the traditional time series model has a simple structure and is mature in application in the financial market,but there are still various limitations.For example,the ARIMA model is a linear model,and the VaR and GARCH family models can measure the nonlinearity of volatility,but the prediction accuracy is low;the neural network model can approximate the nonlinear model arbitrarily,and the performance is better.The BP neural network model is used in many literatures,but the training takes a long time and there is the possibility of generating local minimum values.RBF neural network can overcome the above two shortcomings,and has higher accuracy of function approximation.This paper will establish the ARIMA-RBF neural network model,conduct relevant research on the domestic carbon financial trading market,integrate the advantages of the two models,and effectively overcome the problems in previous research.On the one hand,it expands the financial time series model in carbon emissions trading.The application of the income sequence,on the other hand,enriches the research methods of the carbon financial market,and finally uses a variety of evaluation methods to test the model effect and improve the comprehensiveness of the model evaluation.This paper firstly summarizes the development status of my country’s carbon financial trading market,and takes the average transaction price sequence and yield sequence of the national carbon market and eight pilot carbon markets as the research objects,and expounds the operating characteristics of my country’s carbon financial trading market.Secondly,taking the national carbon market and the return rate sequence of carbon emission rights exchanges in Shanghai,Hubei,Guangdong,Beijing,and Fujian as the research objects,the ARIMA-RBF neural network model is established,which combines the advantages of the ARIMA model and the RBF neural network model.Then use a variety of evaluation methods to compare it with the traditional ARIMA model and ARIMA-GARCH model.The prediction accuracy is higher,and the ARIMA-RBF neural network model is suitable for predicting the national carbon market return rate series.Finally,according to the current situation and modeling results of my country’s carbon financial trading market,several suggestions are put forward for the normal operation and stable development of the national carbon financial trading market.