Categorial economic policy uncertainty indices or Twitter-based uncertainty indices? Evidence from Chinese stock market

Abstract

This paper mainly investigates the performances of Chinese categorial economic policy uncertainty (EPU) indices and categorial Twitter-based uncertainty for predicting Chinese stock market volatility. Results show both types of uncertainty indices can predict Chinese stock market volatility, especially the categorial Twitter-based uncertainty indices, showing uncertainty indices constructed based on social media contain more valuable information than newspaper-oriented uncertainty indices. In addition, we highlight the predictive performances of the Least absolute shrinkage and selection operator (LASSO) model with regime switching for forecasting Chinese stock market volatility.