The term herd behavior derives from the observation that animals that form part of a group sometimes mimic actions of either a leader or each other apparently without thought. In economics and financial markets, this term describes situations in which a large number of agents appear to be making similar decisions. Spurious herding results when decisions are similar, but based on independent analysis by agents. True herding results when decisions are based only on public information about the behavior of others, and private information is neglected. Herd behavior can have a detrimental effect on economic efficiency. It stands in contradiction to the basic tenets of efficient markets and to the claim of the superiority of the market system; it undermines the reliance on rationality that requires that agents base their decisions on all information. Modeling of herd behavior has only recently begun to shed light on its motivations and consequences.
A form of herd behavior results when information cascades develop in markets characterized by the high cost of information collection. When information collection is costly, an agent may make the same decision as previous agents and not seek any private information. The situation is exacerbated by the presence of moral hazards.
Portfolio managers may herd when their performance is evaluated relative to the market, and herding behavior can also be observed in the international lending decisions of banks (see the discussion of international lending below). Agents may be forced to herd in situations where market sentiment will dictate outcome—regardless of the underlying economics. What is an investor to do when the market sentiment is that the stock market will rise? The sentiment itself will cause the market to rise—an individual investor who, perhaps rightly, believes that a bubble is forming cannot expect to gain from this information unless the timing of the collapse of the bubble can be predicted. The rational strategy for this investor would be to herd.
In the public policy arena, public support for or opposition to particular policy measures frequently derives from what are called availability cascades : the more available information is, the more it is considered reliable. Public opposition to smoking is often cited as an example of herding caused by availability cascades.
Herding may be rational behavior in the presence of external threats. Wild animals move together when threatened by a predator—in such cases, the optimal behavior for an individual member is to stay with the group. What is observed as herd behavior is really a consequence of individuals acting rationally.
Less rationally, social pressure may cause individuals to conform to what others are doing. Fads, manias, and fashions may be the consequences of herding resulting from social pressure. Individuals may also seek to safeguard their personal reputation by following the crowd rather than risk the consequences of standing out. Adolescents’ behavior in high schools and conformity to politically correct language are examples of this latter type of herding.
International bank loans to developing countries from 1973 to 1982 provide a good illustration of how information cascades, managerial incentives, and moral hazard lead to herding. The oil price shocks of the 1970s left commercial banks from industrialized countries with large pools of recycled petrodollars and the developing countries with large current account deficits. Although the loan of excess bank liquidity to the developing countries would solve the imbalances for both groups at least temporarily, banks had limited knowledge regarding the credit risks of such loans. Some of the largest banks invested resources in developing the technology for risk assessment, but the large majority of the banks went along for the ride based solely on their observation that the larger banks were willing to lend their own funds to developing countries and were lead-managing multibillion-dollar syndicates to provide funds to these borrowers. For the smaller banks, the marginal cost of doing their own analysis was much higher than the cost of gathering public information. The banks also felt pressure from the stock markets, which compared a bank’s earnings against those of others that were making what had become highly profitable loans (at least until the middle of 1982). The situation was not helped by the blessings given by the governments of Organisation for Economic Co-Operation and Development member states (who believed that recycled funds would support their exporters) and by international development organizations like the International Monetary Fund (which were unable to finance the developing countries’ deficits on their own). Such blessings were taken as an implicit understanding that these governments would help the banks should the banks’ lending behavior get them in trouble. Large commercial banks relied on an implicit protection inherent in their market power, believing that governments would not allow major banks to fail. Smaller banks may have herded for two reasons: (1) they could not replicate the information-gathering and analytic capabilities of the larger banks; and (2) they were sure that central banking authorities could not allow smaller banks to fail while saving the larger ones. Indeed all banks, large as well as small, were propped up after the Third World debt crisis began to unfold in August of 1982.
Herd behavior may be at least partly responsible for the Mexican crisis of 1994 and the East Asian crisis of 1997–1998. In both these situations, investment managers may have continued to invest in assets that had become excessively risky under the belief that markets would have penalized them if they failed to earn returns similar to their competitors.
Herd behavior, by encouraging overinvestment when things look good and mass exodus when that impression is reversed, adds to market volatility. It has, however, been difficult to develop tests that can establish either the presence or the impact of herding with some certainty. Future research in this field will have to focus on isolating real herding from mere similarity of outcomes.
SEE ALSO Banking; Bubbles; Conformity; Developing Countries; Economic Crises; International Monetary Fund; Loan Pushing; Loans; Manias; Moral Hazard; Panics; Rationality; Stock Exchanges
Bikhchandani, Sushil, David Hirschleifer, and Ivo Welsh. 1992. A Theory of Fads, Fashions, Custom, and Cultural Change as Information Cascades. Journal of Political Economy 100: 992–1026.
Bikhchandani, Sushil, and Sunil Sharma. 2001. Herd Behavior in Financial Markets. IMF Staff Papers 47 (3): 279–310.
Darity, William, Jr., and Bobbie L. Horn. 1988. The Loan Pushers: The Role of Commercial Banks in the International Debt Crisis. Cambridge, MA: Ballinger Publishing Company.
Jain, Arvind K., and Satyadev Gupta. 1987. Some Evidence on ‘Herding’ Behavior by U.S. Banks. Journal of Money, Credit, and Banking 19 (1): 78–89.
Arvind K. Jain