A Simple Model of Herding and Contrarian Behaviour with Biased Informed Traders

This paper unveils novel insights into the impact of trader bias on herding and contrarian behaviour in financial markets. We modify the behaviour of informed traders in a sequential trading microstructure model using cumulative prospect theory (CPT). By introducing CPT gain-loss asymmetry and loss-tolerant traders, we formulate a generalized CPT trader herding model. Our findings reveal that in markets with a substantial proportion of loss-tolerant agents, the elimination of gain-loss asymmetry can incur significant costs, emphasizing the necessity of employing the generalized model. Conversely, in markets dominated by loss-averse traders, such an assumption is less costly, allowing reliance on an extended model for closedform results. We establish theoretical upper bounds on loss attitude, determining the threshold that triggers herding and contrarianism, thus facilitating regulatory monitoring. Contrary to previous models, our generalized approach suggests that a trader can engage in both herding and contrarian behaviour rather than a clear-cut preference, and they can occur at mild price deviations instead of only at extreme prices. We reconcile previous experimental evidence, addressing strong contrarian but weak herding tendencies under various market specifications.