The Impacts Of Investor Sentiment And Investor Trading Behavior On ETF Pricing

As ETFs have grown substantially in diversity,market significance and size,there is increased interest by investors and practitioners in the trading of these investment vehicles.Traditional capital asset pricing theory holds that arbitrage activities of rational investors will eliminate asset mispricing and push asset prices close to fundamental values.However,a number of studies find that ETF price in secondary market deviate significantly from net asset value.ETF is considered to be price inefficient as arbitrage is not being effectively executed to drive prices back to their true value.Moreover,frequent price volatility is a feature of ETF market.Based on behavior finance theory,we adhere to the idea of Shiller(2014)'s opinion that researches in light of actual human behavior should take account of how people really think and act.Therefore,this paper constructs fund-level investor sentiment indicator and investor trading behavior indicator to examine how investor sentiment and investor trading behavior affect ETF prices in A-share stock market from the perspective of irrational influence factors.In general,this paper includes the following aspects.First,this paper uses panel data regression model to analyze the impacts of investor sentiment and investor trading behavior on ETF pricing efficiency.Panel regression results show that both investor sentiment and investor trading behavior have significantly positive influences on ETF discount/premium.The combined effects of investor sentiment and investor trading behavior on ETF discount/premium are significantly positive.Furthermore,the above results are still the same after considering the effects of Fama-French three factors and institutional ownership.Second,this paper analyzes the effect of investor sentiment on ETF price volatility.The panel data regressive analysis reveals that investor sentiment have asymmetric influence on ETF price volatility in different periods of sentiment state.Volatility increases with investor sentiment in the periods of high-sentiment state,and decreases with investor sentiment in the periods of low-sentiment state.The magnitude of bullish(bearish)changes in investor sentiment leads to upward(downward)revisions in volatility.Furthermore,the panel quantile regression results exhibit nonlinear patterns across the quantiles obviously: weaker sentiment effects for lower quantiles and stronger sentiment effects for higher quantiles in general.Third,this paper studies the impact of investor trading behavior on ETF price volatility.Specifically,volatility increases with investor trading behavior in the periods of positive buysell imbalance,and decreases with investor trading behavior in the periods of negative buy-sell imbalance.The panel quantile regression results exhibit nonlinear patterns across the quantiles: weaker trading behavior effects for lower quantiles and stronger trading behavior effects for higher quantiles in general.Furthermore,we decompose investor trading behavior into expected and unexpected components and find that their effects on volatility are different.Fourth,this paper analyzes the combined effects of investor sentiment and investor trading behavior on ETF price volatility.Panel regression results show that the combined influences of investor sentiment and investor trading behavior in the same direction on ETF price volatility show consistent directions.Volatility will increase with investor sentiment and investor trading behavior in the periods of high-sentiment state and positive buy-sell imbalance,and vice versa.On the contrary,the combined influences of investor sentiment and investor trading behavior in the different direction on ETF price volatility show inconsistent directions.Volatility will increase with investor sentiment and decrease with investor trading behavior in the periods of high-sentiment state and negative buy-sell imbalance.Volatility will decrease with investor sentiment and increase with investor trading behavior in the periods of low-sentiment state and positive buy-sell imbalance.Panel quantile regression results show that investor sentiment and investor trading behavior have significant combined effects on ETF price volatility.In general,among four states formed by the combination of investor sentiment and investor trading behavior,for higher quantiles,optimistic sentiment has significantly positive effect on volatility and pessimistic sentiment has significantly negative effect on volatility.Similarly,for higher quantiles,positive buy-sell imbalance has significantly positive effect on volatility and negative buy-sell imbalance has significantly negative effect on volatility.Finally,this paper examines the influences of investor sentiment and investor trading behavior on ETF risk-return relation.We sort all ETFs by investor sentiment and ETF price volatility and find: for the ETFs with the largest investor sentiment index,risk-return relation is positive;for the ETFs with the least investor sentiment index,risk-return relation is negative.We sort all ETFs by investor trading behavior and ETF price volatility and find: for the ETFs with the largest investor trading behavior index,risk-return relation is positive;for the ETFs with the least investor trading behavior index,risk-return relation is negative.Furthermore,this paper sorts all ETFs by investor sentiment,investor trading behavior and ETF price volatility and find: for the ETFs with the largest investor sentiment and investor trading behavior indices,risk-return relation is positive;for the ETFs with the least investor sentiment and investor trading behavior indices,risk-return relation is negative.The same direction of investor sentiment and investor trading behavior have superimposed and enhanced effects on risk-return relation.The risk-return relation is uncertain when investor sentiment and investor trading behavior indicators are in the extremely opposite direction.The final joint effect depends on the relative influence of the two irrational factors on risk-return relation.Overall,these results reveal that both investor sentiment and investor trading behavior have significant impacts on ETF pricing.Especially,the empirical results are consistent across two conditional variance models in the study of ETF price volatility and risk-return relation.