The purpose of this paper is to investigate the impact of individual investor sentiment on the return process and conditional volatility of three main US market indices (Dow Jones Industrial Average, S&P500 and Nasdaq100). Individual investor sentiment is measured by aggregate money flows in and out of domestically oriented US mutual funds.
A generalised autoregressive conditional heteroscedasticity (GARCH)‐in‐mean specification is used, where our measure for individual sentiment enters the mean and conditional volatility equation.
For a sample period of six years (February 1998 until December 2004), we find that sentiment has a significant and asymmetric impact on volatility, increasing it more when sentiment is bearish. Using terminology of De Long et al., we find evidence for the “hold more” effect, which states that when noise traders hold more of the asset, they also see their returns increase, and the “create space” effect, which states that noise traders are rewarded for the additional risk they generate themselves.
In contrast to existing studies using explicit measures of market sentiment on low sampling frequencies, the use of daily mutual flow data offers a unique picture on investors' portfolio rebalancing and trading behavior. We propose an integrated framework that jointly tests for the effects of mutual fund flows on stock return and conditional volatility.
