Forecast on Shanghai Composite Index linked with Investor Sentiment Effect

Abstract

Investor sentiment is an important factor that affects investors' decision-making behaviors. Especially when the emotions are very social, people's behaviors will tend to be consistent, leading to market fluctuations. Some scholars tried to study the impact of investor sentiment on market return and volatility. However, they are not able to get a consistent result. This paper constructs an investor sentiment index (CICSI) by principal component analysis. Based on heterogeneous autoregressive (HAR) theory, this paper establishes three HAR models extended by CICSI to forecast the volatility of Shanghai Composite Index. The empirical results reveal that new models’ accuracy is higher than the original one. Data indicates that the decomposed CICSI contains much forecasting information on market volatility, especially in the short-term. By decomposing CICSI, the goodness of fit of the model was improved by 11.08%. This study fills in the gap of previous research by using high-frequency data and decompose investor sentiment. Further study can be applied to find more relative variables to extend the model and improve prediction accuracy.