Stochastic Herding by Institutional Investment Managers
This paper demonstrates that the behavior of institutional investors around the downturn of the U.S. equity markets in 2007 is consistent with stochastic herding in attempts to time the market. We consider a model of large number of institutional investment managers who simultaneously decide whether to remain invested in an assets or liquidate their positions. Each fund manager receives imperfect information about the market’s ability to supply liquidity and chooses whether or not to sell the security based on her private information as well as the actions of others. Due to feedback effects the equilibrium is stochastic and the “aggregate action” is characterized by a power-law probability distribution with exponential truncation predicting occasional “explosive” sell-out events. We examine highly disaggregated institutional ownership data of publicly traded stocks to find that stochastic herding explains the underlying data generating mechanism. Furthermore, consistent with market-timing considerations, the distribution parameter measuring the degree of herding rises sharply immediately prior the sell-out phase. The sell-out phase is consistent with the transition from subcritical to supercritical phase, whereby the system swings sharply to a new equilibrium. Specifically, exponential truncation vanishes as the distribution of fund manager actions becomes centered around the same action – all sell.
Many apparent violations of the efficient market hypothesis, such as bubbles, crashes and “fat tails” in the distribution of returns have been attributed to the tendency of investors to herd. Particularly, in a situation where traders may have private information related to the payoff of a financial assets their individual actions may trigger a cascade of similar actions by other traders. While the mechanism of a chain reaction through information revelation can potentially explain a number of stylized facts in finance, such behavior remains notoriously difficult to identify empirically. This is partly because many theoretical underpinnings of herding, such as informational asymmetry, are unobservable and partly because the complex agent-based models of herding do not yield closedform solutions to be used for direct econometric tests.