Behavioural profiling: A learning-augmented approach to pricing risk

Abstract

To imagine that asset pricing is not dependant on a complex form of behavioural heuristics and interactive game theory it is requisite that we reduce the definition of the participants to that of traditionally defined utility maximising risk-averse uniform automata. This study tackles this statement directly through an application of behavioural theory which speaks to the individual ability of investors to perceive risk, as well as the interactive effects of game theory to distort the perception of risk from exogenous variables to that of endogenous probability beliefs. The result is an asset pricing model which tracks the evolution of investor probability beliefs as traders learn to adapt to their market position captured through the application of a Kalman filter. The behaviourally inspired asset pricing model shows marked improvement over a traditional OLS CAPM in light of evidence that the volatility of equity returns vary over time, finding that estimated pricing errors are reduced by as much as 41%. In comparisons of various behavioural designs, evidence presented suggests that investors tend to price risk towards long run-risk whilst being notably influenced by exposure to lagged market performance. Together, these findings lend support to the hypothesis that investors tend to price risk as a dynamic learning process which is informed by both internal behavioural heuristics of cognisance, as well as position in the market place as a matter of game theory placement. The findings of this study provide a strong basis for the further development of asset pricing in the state-space modelling environment in order to build on a theory of asset pricing through risk perception as a behaviourally dependent self-organised dynamic system.