Research On The Carbon Financial Asset Pricing Considering The Contagion Of High-order Moment Attribute

The rapid increase in greenhouse gas emissions is a direct cause of the negative externality of the global environment.Giving the carbon emission rights to commodity attributes,relying on market-oriented financial measures to solve carbon emission reduction issues have become the main means for international community to respond to climate change and suppress the greenhouse gas emissions.The establishment of carbon financial market is based on the conventions and agreements of the international community in fulfilling its emission reduction responsibilities,and it is developed in the policies and measures of countries to promote the operation of the carbon financial market.As the core of the carbon financial market,an effective carbon pricing mechanism will promote the maturity and perfection of the carbon financial market mechanism,improve the market efficiency,and better serve the implementation of carbon emission reduction.The research of carbon financial asset pricing should not only follow the basic pricing methods of general financial assets,but also reflect the special driving characteristics of carbon price.However,the existing research on carbon pricing focus on the price information transmission and risk volatility spillover of carbon financial assets from the perspective of low-order moment of return,while neglects to study the impact of market asymmetric information and extreme impact on the return from the view of the high-order moment attribute.Especially,with the enhancement of global capital flow,the relation between carbon financial market,capital market and energy market have been confirmed to have a low-order moment attribute information linkage.In addition,there will also produce the risk contagion relation in the perspective of the high-order moment attribute such as skewness and kurtosis caused by asymmetric information and extreme event impact in the market.The significant risk contagion in the view of the high-order moment attribute can provide an effective explanation for the price changes caused by irrational coordinated movements and extreme shocks in the market based on the theory of high-order moment risk contagion.This explanation perspective is consistent with the special characteristics of carbon financial assets,such as the feature of market asymmetry and extreme shock sensitivity.Therefore,the risk contagion in the view of the high-order moment attribute is included in the pricing framework of carbon financial assets,which conforms to the special characteristics of carbon assets,and can explain the premiumof carbon financial assets from a new evidence factor.Based on this,the innovation and conclusion of this research are as follows:(1)Constructed the pricing framework of carbon financial assets that considering the risk contagion of high-order moment attribute.Firstly,this research extends the CAPM to multi-factors framework,and then a mult-factors pricing framework for carbon financial assets is formed.Secondly,based on the high-order moment attribute risk contagion theory of carbon financial assets,this research tests and identifies the risk contagion relationship between carbon price and its pricing factors in the pricing framework,studies the risk contagion relationship and contagion intensity under the heterogeneity of market volatility trend,and explains the premium of the carbon financial assets from the perspective of market irrational co-movement and extreme impact.Thirdly,the identified high-order moment risk contagion relationship is fed back to the extended high-order moment carbon financial asset pricing framework to form a novel pricing framework that considering the high-order moment attribute risk contagion.The proposed theoretical framework follows the idea of "general to special",and integrates the high-order moment risk contagion with the special characteristics of carbon assets into the pricing framework,provides new evidence for explaining the carbon financial assets premium.(2)The nonparametric statistical econometric model is used to test the contagion relationship between carbon financial assets and their pricing factors based on the heterogeneity of volatility trend.The fluctuation of carbon market implies the information reflection of carbon financial assets to the extreme shock or asymmetric shock.The research of the risk contagion based on the heterogeneity of volatility trend conforms to the special characteristics of carbon financial market fluctuation.The results show that:(1)There is not only a risk contagion relationship between carbon financial assets and their pricing factors in the channel of low-order moment,but also a significantly contagion relationship in high-order moment channels,such as the channel of co-skewness,co-kurtosis and co-volatility.This shows that the risk contagion relationship between carbon financial assets and their pricing factors caused by market asymmetric information and extreme impact has become a new factor evidence that affects the price mechanism of carbon assets.(2)The risk contagion intensity under the trend of rapid fluctuation is greater than that of slow fluctuation.The difference of contagionintensity in different risk fluctuation trends is a representation of the carbon market return and the corresponding relation between risk and return.This conclusion provides an analysis basis for screening the carbon pricing factors with risk contagion relationship in the high-order moment channel,and also help to analyse the pricing model.(3)Constructed a multi-layer and multi-variable LSTM model to realize the fitting and prediction of the proposed carbon pricing framework.Based on the identified carbon returns and their pricing factors with the significantly risk contagion relationship in high-order moment attribute,this research constructs an empirical model based on the proposed carbon pricing framework,and employ the machine learning model of multi-layer multi-variable Long-Short Term Memory neural network(Multi-LSTM)fitting the proposed pricing framework.In the experiment,this research determines the optimal network structure and parameters through experimental means for improving the performance of model fitting and achieving effective convergence.The study found that the proposed pricing frameworks that consider the high-order moment attribute risk contagion can perform better for fitting and forecasting the carbon returns,especially in the longer period of12 months,that compared to the pricing frameworks that do not consider the contagion of high-order moment attributes.Furthermore,the prediction performance and stability of the Multi-LSTM model is significantly better than other deep network models(Multi-GRU,RNN,MLP),volatility model(Garch-M)and neural network model(BP).The conclusion confirm the superiority of the proposed pricing framework and the fitting advantages of the machine learning method.In theory,the contribution of this research proves that the proposed pricing framework of carbon financial assets considering the risk contagion of high-order moment attributes can provide a powerful explanation for the premium of carbon assets,shows the rationality and effectiveness of incorporating the risk contagion of high-order moment attributes into the pricing framework,and develops the pricing theory and method of carbon assets.In practice,the conclusion is helpful to promote the maturity and perfection of carbon asset pricing mechanism,promote the realization of emission reduction goals,provide reference for the investment and financing decision-making of carbon financial market in the uncertain environment.