Is the Fama and French three factor model robust to the pricing of risk preferences?
摘要
It is already well established that neither the bias inherent in, nor the extent to which a multifactor specification of the Capital Asset Pricing Model (CAPM) prices behaviors (priors), as opposed to rationality (risk preferences) is decipherable from the output of any such model. Hitherto, empirical structures that facilitate either inference have been lacking. This study’s formal theory directly spans an empirical structure which nests inferences in respect of the bias inherent in a CAPM in inferences in respect of the extent to which the same CAPM prices risk preferences, as opposed to priors that are systemically realized (‘systemic priors’) within populations of agents. The resulting empirical evidence in respect of the Fama and French three factor model is unequivocal, namely that it solely prices systemic priors, not agents’ risk preferences, a finding that is consistent with the suspicion, in prior studies, that the market betas that are generated by the model are so high, it is unlikely they evince rational pricing. Contrarily, conditional on the presence, in markets, of non-globally risk averse agents, a multifactor CAPM which incorporates the market portfolio and market skewness factors prices agents’ risk preferences and systemic priors and outperforms all other candidate models.