Herding is a deliberate decision to imitate the actions of others. In financial markets with private information herding can be efficient for an individual asset manager, but increases the risks that the market as a whole is inefficient and fragile, particularly in the case of “information cascades”. A paper of Michael McAleer and Kim Randalj provides empirical evidence of herding in a range of futures markets.
“Herding, Information Cascades and Volatility Spillovers in Futures Markets”, Michael McAleer and Kim Randalj
Journal of Reviews on Global Economics, 2013, vol. 2, pages 307-329
The below are excerpts from the paper. Emphasis and cursive text have been added.
What is herding?
“Herding is defined as a conscious decision by agents to mimic the actions of others. Thus, herding is a deliberate decision, which should not be confused with correlated decision making that is purely incidental…In the presence of uncertainty, mimicking the decision of others may be perfectly rational, so that herding need not be associated with irrational behaviour. However, the herding equilibrium may not be socially efficient and prices may be more volatile than if agents had acted independently of each other…Herding is most likely to occur in situations where the decisions of others are observable, as it is not possible to copy what cannot be observed. ”
What explains herding?
“Becker’s (1991) restaurant example shows that, when confronted with different choices and without private information, agents may base their decisions upon the number of patrons. This outcome may arise because previous patrons are believed to have made their decisions based upon private signals regarding quality.”
“[In financial markets] herding may arise for a variety of reasons. Managerial remuneration often depends upon reputation, but the principals may be uncertain of managerial quality. Thus, poor managers have an incentive to copy the decisions of other managers in order to mask their inferiority… Agents may also be compensated according to performance relative to their peers. In this instance, risk-averse managers will be unlikely to deviate from their peers, and will tend to cluster in their portfolio decisions.”
“The type of herding most directly related to the context of futures market traders is based on the theory of information cascades… An information cascade arises when decisions are made by each agent sequentially, but agents begin to ignore their private signals in favour of the observed actions of previous agents. These signals are generally either ‘good’ or ‘bad’ and, importantly, agents cannot observe the signals received by other agents. However, the probability of the good or bad signal can be inferred from the actions of others, which are assumed to be observable. As agents are unsure of the quality of the signals, the actions of others are used to update their beliefs about, for example, the true value of an asset.”
What is the problem with herding?
“Herding is more likely, the less certain is an individual of the private signal. Moreover, if the signals received are noisy, and hence provide little certainty, the probability that the herd arrives at an incorrect decision (such as failing to invest when they should) will be high. This illustrates that herding can lead to inefficient social outcomes, despite the fact that agents have acted in a self-interested and rational manner. Thus, in situations where agents are prone to herd, the market mechanism cannot be used to reach a socially efficient outcome.”
“Furthermore, the theoretical models illustrate that [information] cascades are fragile. Cascades imply that prices reflect only a narrow information set, so the arrival of new information can lead agents to re-evaluate their choices and cause the cascade to shatter… The fragility of information cascades has important consequences for financial market stability. If the herd is prone to alter its positions, then herding will cause financial markets to be excessively volatile.”
What is the evidence of herding in financial markets?
“A primary motivation for this paper is to test for herding among futures market participants, namely small traders, which is measured by including the lagged positions of large futures speculators… This paper uses futures position data in nine different markets of the Commodity Futures Trading Commission (CFTC)…Large speculators are assumed to be informed traders because their size allows them to acquire high quality information. Moreover, there is an incentive for large speculators to invest considerable time and resources in evaluating the information at their disposal…The empirical model is dynamic and is estimated with a lag length of one, as weekly data are used and higher-order lags quickly become dated because news arrives continuously.”
“Evidence consistent with herding among small traders is found for the Canadian dollar, British pound, gold, S&P 500 and Nikkei 225 futures. Consistent with survey-based results on technical analysis, the positions are significantly correlated with both current and past market returns.”
“This pattern of trading behaviour is consistent with expectations as to how small noise traders formulate their trading strategies. Being small, they do not have the resources to acquire costly information, and hence tend to rely on price movements to determine their investments. The coefficients of lagged spot returns were statistically significant in the currency and oil markets, suggesting that small traders may cause currency markets to be unstable.”