Vous êtes sur la page 1sur 3

Research proposal

Mispricing and heterogeneous agents in financial markets: a coordination problem

Evidence in financial markets, about aggregate stock market and cross-section of average
returns, and individual trading behaviours seem to be not easily understood in the traditional
finance paradigm. A more accurate description of financial markets is thus needed to account for
such evidence.
First, financial markets may exhibit in some occasions mispricing – asset prices differ from their
fundamental values. Even if assessing fundamental values of financial assets is a tricky matter,
some features of asset prices are most plausibly interpreted as deviations from fundamental
values, which persisted over time. Persistent mispricing will result in the predictability of asset
returns, from which some investors can take advantage and systematically make profits. But this
would invalidate the predictions of traditional finance which suggest that no investor can
systematically beat the market and earn expected profits in excess of equilibrium levels. Hence
evidence of mispricing in financial markets, for which no rational explanations exist, cannot be
fully explained by traditional finance founded on the Efficient Market Hypothesis (EMH) which
assumes that “prices are right” – asset prices reflect its fundamental value 1. However since
financial markets are concerned with allocating capital to the most promising investment
opportunities, “prices are right” becomes a central issue.
Second, agents that participate in financial markets – investors – are heterogeneous – institutional
versus individual investors, differences in endowments, in trading strategies or with respect to
information – and may exhibit less-than-perfectly rational behaviours. Thus in such context,
dealing with a representative agent, assumed fully rational2, may not be reasonable. One may then
ask how to account for behavioural complexities? How the latter may affect investors and their
trading? But also how the interactions between heterogeneous agents may influence asset prices?
In such context, I believe that one way to go is to inspect whether mispricing phenomena –
including bubbles as an extreme case – are due to a coordination problem between heterogeneous
agents. For that a better understanding of agents’ behaviour is needed, in order to investigate how
investors interact in the market and what may be the outcome of their interactions. From this point
of view, I believe that results in behavioural finance may be enlightening.

In a preliminary work, I investigated first some features of asset prices considered as


anomalies by traditional finance in order to consider why alternative explanations are needed.
Actually if one believes in efficient markets, one would believe that market prices provide the
information that investors need to trade in an optimal way. But then one may ask whether this
assumption is reasonable, whether investors are actually able to read properly market prices and
use market information. The examination of some financial anomalies – such as changes to the
S&P 500 Index, international investing and home bias, pricing of closed-end funds – provides
some elements to believe that this may not be the case. Indeed first, all investors may not have the
expertise and the knowledge to identify and trade some mispricings. Second, not only new
information but also investors’ sentiment seems to influence asset prices.
Furthermore one who believes in efficient markets will disregard the current market, since broad
diversification is the ideal goal to be achieved. But such belief may lead investors to make
considerable mistakes when they trade because it prevents them to account for the trading activity
of the others. Whenever rational explanations fail to account for financial anomalies, alternative
explanations in terms of investors’ expectations and investors’ awareness may be noteworthy to

1
Fundamental value is generally assumed to be the discounted sum of expected future cash flows.
2
Rationality implies two ideas. First, when they receive new information, agents update their beliefs correctly, in
the manner described by Bayes law. Second, given their preferences, agents make choices that are normatively
acceptable, i.e. consistent with Savage’s notion of Subjective Expected Utility (SEU).
understand asset pricing. Thus looking more closely at investors’ behaviour may be enlightening
to understand market behaviour and some of these anomalies. How heterogeneous market
participants make their decisions, use market information and trade is then a central issue.
Consequently I considered also some works in behavioural finance since this approach
proposes a more accurate explanation of agents’ behaviour. I believe that this is a necessary first
step to understand how investors choose their portfolios and how they trade. Actually while
traditional finance seeks to understand asset pricing through models in which agents are rational
and exhibit consistent beliefs3, behavioural finance proposes some crucial elements to understand
how agents may deviate from rationality and proposes to explain some financial anomalies
through models in which agents are no more fully rational. Indeed behavioural economists turn to
the extensive experimental evidence compiled by cognitive psychologists on the biases that arise
when agents form their beliefs, and on agents’ preferences. This literature widely documents
human patterns of less-than-rational behaviours. Facing these results, it may be tricky to sustain
that agents are able to read and use perfectly market information. First agents may be subject to
judgmental biases that prevent them to form correct beliefs and to use properly the available
information. Moreover agents may make choices that are not normatively acceptable and may not
be able to achieve their goals in the best way. While there is no evidence supporting clear
irrationality among investors, human patterns of less-than-rational behaviours may be noteworthy
to explain financial market behaviour, and why in financial markets, investors’ demand for
financial assets may be influenced by their beliefs or sentiments, which may lead them to trade not
only based on market information.
These considerations on agents’ rationality are instructive to understand asset pricing, since they
provide elements to understand how agents use and interpret information – useful to define
accurately agents heterogeneity – and how agents make their decisions.
Behavioural finance rests on a second building block4: limits of arbitrage. Contrary to traditional
finance, market forces which bring back asset prices toward fundamental values may be limited 5.
In a market where rational investors interact, less-than-perfectly rational behaviours can have a
substantial and long-lived influence on prices. This is instructive when one wants to account for
the influence of investors’ sentiment on asset prices.
Thus behavioural finance, providing a more realistic analysis of agents behaviours and asset
pricing, helps understand how agents, not fully rational, behave, how they interact in the market
and what is the outcome of their interaction. I believe that the results in behavioural finance have
to be considered in order to propose an alternative explanation of mispricing phenomena in
financial markets based on a coordination problem between heterogeneous agents.
Finally I examined some works in which models based on heterogeneous agents are used
in order to explain some mispricings. Actually heterogeneity of agents is widely accepted in
economics and finance literature as the explanation for trade and is increasingly integrated as a
key feature of asset pricing models. This investigation gives support to our prior intuition about
the relevance of agents’ heterogeneity in explaining financial anomalies, and about the diversity in
the sources of heterogeneity – differences in preferences, differences in endowments, and
institutional differences6 between investors on financial markets. However the analysis sheds light
on a more crucial source of heterogeneity based on information. In asset pricing models, this
feature may arise because of two assumptions: differential information and differential
interpretation. In the first case, there is asymmetric information between a group of agents that
observe a private signal, and the rest of the population that has to learn the fundamental value
from public information such as prices. Asymmetric information causes heterogeneous
expectations among agents which play a significant role in asset pricing. The second assumption

3
Consistent beliefs express the fact that agents’ beliefs are correct.
4
The idea that behavioural finance is built on two pillars - limits of arbitrage and investors’ psychology - is
originally due to Shleifer and Summers (1990).
5
DeLong et al. (1990), Shleifer and Vishny (1997), Wurgler and Zhuravskaya (2002).
6
Institutional versus individual investors.
is based on the fact that public information can be interpreted in different ways by agents. Agents
use different “models of the market” to update their subjective valuation, which might lead them
to hold different beliefs.
Furthermore as suggested by Kandel and Pearson (1995)7, I believe that the assumption that
agents interpret public information identically is too restrictive. Relaxing this assumption would
change significantly the analysis and may help provide more accurate explanations for financial
anomalies. These elements give support to our prior intuition about the fact that it could be
controversial to assume that agents are able to read perfectly market prices.
While my previous work has been instructive for clarifying several issues – theoretical
approach to be adopted, relevant features of agents’ heterogeneity and significance for the
analysis of financial anomalies – and convincing enough to believe that mispricing phenomena
may be better understood by focusing on the interactions between heterogeneous agents, some
issues remain unsolved and would be the object of future research.

When standard market imperfections – such as cost of information and trade, capital
constraints and imperfect substitutes – do not help explain mispricing, I believe that a framework
based on synchronisation risk8 offers a more accurate and promising way to propose an alternative
explanation based on a coordination problem between heterogeneous agents. In such framework,
where rational arbitrageurs interact with behavioural agents – boundedly rational – arbitrage is
limited because the former face uncertainty about when other rational arbitrageurs will start
exploiting a common arbitrage opportunity. Synchronization risk stems from the uncertainty about
the market timing decision of other rational arbitrageurs, since they become sequentially aware
about changes in the fundamental value of the risky asset. While rational arbitrageurs have to
coordinate their actions in order to trade effectively the mispricing, they face uncertainty about
when the other rational arbitrageurs will also trade it. A coordination problem arises among
rational arbitrageurs, when some decide to trade the mispricing. Consequently, arbitrage is
delayed and the mispricing may persist in the market even if some rational arbitrageurs are aware
about it and may trade it efficiently.
While this analysis focuses mainly on the decision-making of rational arbitrageurs, it would be
worthwhile focusing also on the decision-making of behavioural agents. One would then provide
a more accurate analysis of the interactions between heterogeneous agents, not only based on
differential information but also on differential interpretation, as one way to relax the assumption
that agents interpret identically market information. In such framework, rational arbitrageurs
would not only have to account for the behaviour of other rational agents but also have to account
for the decisions of behavioural agents.
Modelling explicitly the behavioural agents’ decision-making would be based on well-
documented and relevant judgmental biases. Recent empirical evidence suggests that the use of
internet to trade is more and more widespread among investors and that this new “style investing”
promotes overconfidence. After going online, investors seem to trade more actively and more
speculatively. Thus one psychological bias that may be further explored and worthwhile
incorporating in such model could be overconfidence. Furthermore the increasing use of internet
changes significantly how information is delivered to investors and the ways in which investors
can act on that information.

7
Kandel, E., and Pearson, N., 1995, Differential interpretation of public signals and trade in speculative markets,
Journal of Political Economy 103 :4, 831-872.
8
First proposed by Abreu and Brunnermeier (2002).

Vous aimerez peut-être aussi