Asset Pricing, Signal Type and Overconfident Investors

Models reflecting behavioural biases (eg conservatism, overconfidence) provide rationales for short-term overreactions and long-term mean reversion. We expand the overconfidence framework to generate new testable hypotheses based upon differing reactions to private/public signals. We demonstrate that differing combinations of private/public signals, both in time and through time, lead to differing adjustment processes in terms of underreactions, overreactions and serial autocovariances. The generalised implications are operationalised by using a number of instruments to differentiate between firms with predominately public versus private signals. Testable implications will be empirically investigated using Australian, US and UK data.