Forecast Research on the Fluctuation Risk of Green Financial Market in China based on the GARCH-LSTM Mixed Model

Abstract

With the rapid development of China’s green financial market, there is also the problem of uncertain risk of fluctuation in the transaction price of green financial products. Therefore, it is imperative to give play to the role of price fluctuations in early warning of risks in green financial markets. Based on the closing price data of green stock market and green bond market from April 19, 2018 to April 18, 2023, this paper combines the traditional time series model and deep learning model to fit and forecast the yield volatility of green financial market, and quantifies the risk value of green financial market. The results show that: (1) Both the green stock market and the green bond market have the characteristics of leptokurtosis and fat-tail, volatility aggregation and long-term memory to varying degrees. (2) The volatility risk in the green financial market has obvious differentiation characteristics, and the volatility in the green stock market is significantly higher than that in the green bond market. (3) The prediction accuracy of LSTM long-term and short-term memory model is significantly higher than that of GARCH family model. The GARCH-LSTM mixed model can further improve the prediction accuracy. According to the research results, the combination of modern artificial intelligence algorithm and classical econometric theory can become an important means to identify the volatility risk of green financial market, and put forward meaningful suggestions for the risk management of green financial market.