Noise traders, excess volatility, and a securities transactions tax
Journal of Financial Services Research (1996)
- ISSN: 09208550
- DOI: 10.1007/BF00115671
Available from www.springerlink.com
or
Abstract
Proponents of a securities transactions tax have suggested that such a tax may reduce stock return volatility. The argument is that, to the extent that short-term speculative trading volume is the source of excess volatility, a tax that reduces such volume will reduce volatility. In the context of a simple general equilibrium model, it is shown that this partial equilibrium argument is misleading and in large part incorrect. In the absence of a tax, the model generates equilibria in which the risky asset's price exhibits excess volatility and agents engage in excess trading activity owing to the presence of destabilizing noise traders. Within the context of the model, it is shown that, although a transactions tax can reduce the volatility of the risky asset's price, the reduction in price volatility is accompanied by a fall in the asset's price as agents discount the future tax liability associated with risky asset ownership. Consequently, although price volatility may decrease slightly, the fall in equilibrium prices more than compensates, and the volatiltiy of risky asset returns unambiguously increases with the level of the transactions tax.
Page 1
Noise traders, excess volatility, and a securities transactions tax
Noise Traders*
James Dow
London Business School
and
Gary Gorton
The Wharton School, University of Pennsylvania and NBER
April 25, 2006
Abstract
Noise traders are agents whose theoretical existence has been hypothesized as a way of solving
certain fundamental problems in Financial Economics. We briefly review the literature on noise
traders. The is an entry for The New Palgrave: A Dictionary of Economics, 2nd Edition (Palgrave
Macmillan: New York), edited by Steven N. Durlauf and Lawrence E. Blume, forthcoming in
2008.
* We thank Pete Kyle for comments.
Page 2
1
“Noise traders” are economic agents who trade in security markets for non-information-based
reasons. The existence of noise traders was theoretically posited as a solution to the “no trade” or
“no speculation” results of Grossman and Stiglitz (1980) and Milgrom and Stokey (1982). These
authors showed that it is impossible under most circumstances for an agent with superior
information to profit from that information by trading. The intuition for the “no trade” result is as
follows. A buyer of an asset is prepared to pay a seller a price p only if the buyer believes that
conditional on the seller agreeing to sell the asset, the value of the asset exceeds p. But, then the
seller, knowing this, is at least as well off keeping the asset. So, no one trades.
But, we do observe trade in the world. Moreover, no trade is difficult to reconcile with the notion
of asset market efficiency, in which prices allegedly contain all available information. If some
agents produce costly private information and then trade on their private information, security
prices will reflect some or all of the information and hence become more informationally
efficient. To explain how informed traders can cover the costs of information production when
they trade in securities markets, someone in the market must lose money trading against them.
“Noise traders” or “liquidity traders” are the names given to the traders who lose money, on
average, when they trade. Their trade then provides the subsidy to cover the cost of information
production by the informed traders
The idea that there are traders who systematically lose money trading securities leads to obvious
questions. Do noise traders really exist? Who exactly are noise traders in reality? How do noise
traders survive and persist when they are losing money trading?
Rational Expectations and Efficient Security Markets
In security markets, prices are alleged to reflect “all available information.” But, how does this
come about? What is the information, and how is it aggregated into the price? The concept of a
Rational Expectations Equilibrium (REE) gave formal content to the notion of “market
efficiency,” which has been a central concept in financial economics for thirty years. The idea is
that if agents understand the economy, and understand how markets work, then they know that
current prices reflect the information which is known to some agents, but maybe not to others.
The uninformed agents understand the link between current prices and the information of the
informed agents, and so can infer something about the information in prices. When the prices
that prevail in equilibrium coincide with what the uninformed agents can learn from the prices,
and with the actions taken by the informed agents, who trade on their information knowing that
the uninformed agents will infer (some or all) of the information, then the equilibrium is said to
be a Rational Expectations Equilibrium. The idea that prices can convey information, in the sense
of REE, is due to Lucas (1972).1 Also, see Green (1973) and Radner (1979).
But, when all the information of the informed agents is revealed, in a fully revealing REE, there is
a problem if information acquisition is costly. Grossman (1976) considers a model of the stock
market in which there are two types of traders “informed” and “uninformed.” Informed traders
take positions in the market based on their information. Uninformed traders have no information,
but know that prices will reflect the information of the informed traders. Grossman shows that the
equilibrium prices aggregate and reveal the information perfectly, “but in doing this the price
system eliminates the private incentive for collecting the information” (p. 574). Grossman is quite
clear in identifying the paradox, but he also proposes a solution:
1 Grossman (1981) provides a brief intellectual history of REE. Also, see Allen and Jordan (1998).
“Noise traders” are economic agents who trade in security markets for non-information-based
reasons. The existence of noise traders was theoretically posited as a solution to the “no trade” or
“no speculation” results of Grossman and Stiglitz (1980) and Milgrom and Stokey (1982). These
authors showed that it is impossible under most circumstances for an agent with superior
information to profit from that information by trading. The intuition for the “no trade” result is as
follows. A buyer of an asset is prepared to pay a seller a price p only if the buyer believes that
conditional on the seller agreeing to sell the asset, the value of the asset exceeds p. But, then the
seller, knowing this, is at least as well off keeping the asset. So, no one trades.
But, we do observe trade in the world. Moreover, no trade is difficult to reconcile with the notion
of asset market efficiency, in which prices allegedly contain all available information. If some
agents produce costly private information and then trade on their private information, security
prices will reflect some or all of the information and hence become more informationally
efficient. To explain how informed traders can cover the costs of information production when
they trade in securities markets, someone in the market must lose money trading against them.
“Noise traders” or “liquidity traders” are the names given to the traders who lose money, on
average, when they trade. Their trade then provides the subsidy to cover the cost of information
production by the informed traders
The idea that there are traders who systematically lose money trading securities leads to obvious
questions. Do noise traders really exist? Who exactly are noise traders in reality? How do noise
traders survive and persist when they are losing money trading?
Rational Expectations and Efficient Security Markets
In security markets, prices are alleged to reflect “all available information.” But, how does this
come about? What is the information, and how is it aggregated into the price? The concept of a
Rational Expectations Equilibrium (REE) gave formal content to the notion of “market
efficiency,” which has been a central concept in financial economics for thirty years. The idea is
that if agents understand the economy, and understand how markets work, then they know that
current prices reflect the information which is known to some agents, but maybe not to others.
The uninformed agents understand the link between current prices and the information of the
informed agents, and so can infer something about the information in prices. When the prices
that prevail in equilibrium coincide with what the uninformed agents can learn from the prices,
and with the actions taken by the informed agents, who trade on their information knowing that
the uninformed agents will infer (some or all) of the information, then the equilibrium is said to
be a Rational Expectations Equilibrium. The idea that prices can convey information, in the sense
of REE, is due to Lucas (1972).1 Also, see Green (1973) and Radner (1979).
But, when all the information of the informed agents is revealed, in a fully revealing REE, there is
a problem if information acquisition is costly. Grossman (1976) considers a model of the stock
market in which there are two types of traders “informed” and “uninformed.” Informed traders
take positions in the market based on their information. Uninformed traders have no information,
but know that prices will reflect the information of the informed traders. Grossman shows that the
equilibrium prices aggregate and reveal the information perfectly, “but in doing this the price
system eliminates the private incentive for collecting the information” (p. 574). Grossman is quite
clear in identifying the paradox, but he also proposes a solution:
1 Grossman (1981) provides a brief intellectual history of REE. Also, see Allen and Jordan (1998).
Sign up today - FREE
Mendeley saves you time finding and organizing research. Learn more
- All your research in one place
- Add and import papers easily
- Access it anywhere, anytime
Start using Mendeley in seconds!
Readership Statistics
10 Readers on Mendeley
by Discipline
80% Economics
by Academic Status
30% Assistant Professor
30% Ph.D. Student
20% Student (Master)
by Country
40% United States
20% Germany
10% Italy