Investor sentiment and volatility of exchange-traded funds: Evidence from China

This study examines the relationship between fund-level investor sentiment and return volatility of exchange-traded funds (ETFs). We use the first principle component to form a composite fund-level investor sentiment index that is based on the common variation in four underlying proxies for individual fund sentiment for each ETF: relative strengthen index, psychological line index, Bull and Bear Index and trading volume. And we employ the GARCH model and the EGARCH model to estimate daily ETFs' return volatility, respectively. In our tests, the panel data regressive analysis reveals that investor sentiment reliably predicts ETFs' return volatility in different periods of sentiment state. Specifically, return volatility increases with investor sentiment in the periods of high-sentiment state, and decreases with investor sentiment in the periods of low-sentiment state. Furthermore, the panel quantile regression results exhibit nonlinear patterns across the quantiles obviously: weaker effects for lower quantiles and stronger effects for higher quantiles in general. More importantly, the empirical results are stable across different conditional variance models.