Review of Quantitative Finance and Accounting, 17: 327–350, 2001
2001 Kluwer Academic Publishers. Manufactured in The Netherlands.
Empirical Identiﬁcation of Non-Informational Trades
Using Trading Volume Data
Department of Finance, Bauer College of Business, University of Houston, Houston, Texas 77204-6282
OLIVER M. RUI
Department of Accountancy, Hong Kong Polytechnic University, Hung Hom, KowLoon, Hong Kong
Abstract. This paper empirically identiﬁes non-informational and informational trades using stock returns and
trading volume data of the U.S., Japanese, and U.K. stock markets and ﬁve individual ﬁrms. We achieve the
identiﬁcation by imposing a restriction from theoretical considerations. Our results show that trading volume is
mainly driven by non-informational trades, while stock price movements are primarily driven by informational
trades. We also ﬁnd that, around the 1987 stock market crash, trading volumes due to non-informational trades
increased dramatically, while the decline in stock market prices was due mainly to informational trades. Increases
in volatilities both in returns and in trading volumes during and after the crash are mainly due to non-informational
trades. Regarding the trading volume-serial correlation in the stock returns relationship, we ﬁnd evidence that is
consistent with theoretical predictions that non-informational components can account for high trading volume
accompanied by a low serial correlation of stock returns.
Key words: informational trade, non-informational trade, trading volume
JEL Classiﬁcation: C32, G12, G14
Day-to-day movements in stock market prices and expected stock returns may occur for
two reasons. One is due to public information that causes all investors to change their valu-
ation of the stock market because of new information about fundamental shocks affecting
it. The other is non-informational factors, such as interactions among different groups of
investors with heterogeneous information. We call the former informational trade and the
latter non-informational trade in this paper.
Some researchers introduce non-informational
trades into models by exogenously shifting misperceptions of future stock payoffs, by irra-
tional noise trading (e.g., DeLong, et al., 1989, 1990), or by over-conﬁdent investors who
overestimate the precision of their private signal about security values (Daniel, Hirschleifer
and Subrahmanyam, 1998). Others introduce them by heterogeneous information and in-
vestment opportunities or by shifts in the risk aversion of some traders (e.g., Admati and
Pfleiderer, 1988, 1989; Campbell, Grossman and Wang, 1993; Wang, 1994).