Risk Contagion and Price Forecasting of China Carbon Market Based on High Order Moment Attribute

Li, Ni (2025) Risk Contagion and Price Forecasting of China Carbon Market Based on High Order Moment Attribute. PhD thesis, UNIMAS.

	PDF dsva_Ni Li.pdf Restricted to Repository staff only Download (150kB) | Request a copy
	PDF Thesis PhD_Ni Li.pdf Download (4MB)

Abstract

As an effective mechanism for addressing climate issues, the carbon market plays an important role in reducing global greenhouse gas emissions. The core function of the market is achieving emissions reduction target by the manner of market price mechanism. So, the carbon price is the key point. This thesis focuses on studying the China carbon price forecasting and its price driving mechanism after considering the impact of high order risk contagion relationship, which supports a more convinced and innovative evidence for explaining the carbon price formation that is significantly different from previous researches. In terms of the research variables, the carbon market has closely information linkages and spillovers with the capital market, homogeneous product market, and energy market products. In this regard, this thesis uses the daily transaction price of Hubei Carbon Emission Allowances (HBEA) as the representative indicators of China carbon price, takes the daily settlement price of the European Carbon Emission Allowance Futures (EUAF) as the representative variables of carbon homogeneous market, selects the daily trading price of Jiaotan futures (JTF), Jiaomei futures (JMF),and Brent crude oil futures (Oil) as the special variables of energy market, and selects the daily trading price of China security index 300 (CSI300) as the special variables of capital market. In terms of model design and empirical discussion, firstly, this thesis measures the risk contagion relationship between the carbon market and its infected markets based on the idea of Markov theory, then designs a carbon price state transition model to classify the high, medium and low volatility state of the carbon market. Secondly, constructs and tests the risk contagion channels between carbon price and its infected markets, that is the low order moment risk contagion channel of Forbes Rigobon contagion (FR), and the high order moment risk contagion channels of Co-Skewness (CS), Co-Kurtosis (CK) and Co-Volatility (CV). And finally, designs a high order risk contagion carbon price forecasting model (HOC-LSTM) to forecasting the price of China carbon market. The main conclusion of this thesis are as follows: firstly, the high order risk contagion carbon price forecasting framework support a convinced theoretical support for forecasting the China carbon price. Secondly, there is no risk contagion relationship in low order moment channels, but significant risk contagion relationship in high order moment channels no matter the carbon market in rapid and slow change. Thirdly, the HOC-LSTM model constructed in this thesis has a significant superiority in forecasting China carbon price then other comparative models, such as the Gated recurrent unit (GRU), Multi-Layer Perceptron(MLP), Gradient Boosting Decision Tree (GBDT), Extra Trees Regressor (ETR) and Back Propagation Neural Network (BPNN), the high order risk contagion channels are indispensable factors for explaining carbon price formation mechanism. Those results can not only provide reference for investors and emission reduction entities to make investment and financing decisions, analyze price trends, but also contribute to technical references for government departments to promote the construction of carbon market pricing mechanisms and market efficiency. Keywords: China carbon market, risk contagion, price forecasting, high order moment, HOC-LSTM

Actions (For repository members only: login required)

View Item