Behavioral betas and asset pricing

Abstract

The definition of the 2013 Nobel Prize in Economics laureates is directly related to their empirical work in asset pricing. Eugene Fama, Lars Peter Hansen and Robert Shiller conducted research that aimed to improve understanding of asset pricing process. Fama (1970) discussed about how the diffusion of information and its incorporation into prices occurs, virtually eliminating the predictability of returns in the short term. Some years later, Shiller (1981) questions whether changes in stock prices are caused by rational expectations, as the efficient market hypothesis proposed by Fama (1970) or by other factors. As a result, Shiller (1981) showed that stock prices were historically more volatile than expected, assuming a strictly rational market. When comparing the movements in stock prices with the variation in dividends, he showed that volatility in prices was not explained by the rational variation of expected future dividend payments. Later, Hansen (1982) developed the econometric method called Generalized Method of Moments (GMM) that has made possible the estimation considering particular issues of asset returns like endogeneity. In this context, a number of works have been done in order to find a pricing model that explains actual phenomena in the market. The assumption of rationality of agents which is the basis of neoclassical theory cannot satisfactorily explain the facts observed in practice. There is a growing body of theoretical and empirical academic works that examine the role of market sentiment and its implications for both financial markets as well as for companies and investors. Studies attempt to go beyond the mere fact that investor sentiment affects stock prices (BLACK, 1986; LONG et al, 1990; HONG; STEIN, 1999; BAKER; WURGLER, 2006; FRAZZINI; LAMONT, 2008) but they propose alternative ways to measure this difficult-to-capture concept and verify their relationship with stock prices. Therefore, I intend to analyze whether there are stocks of some companies that may be more affected by changes in market sentiment that others, considering firm characteristics - for instance, attributes related to size, age, industry or even profitability. These attributes may be related to hard-to-evaluate companies, making them more susceptible to changes in the level of market sentiment. The study will include the analysis of the Brazilian and U.S. market data in order to identify whether there are relevant differences in the results when comparing developed versus developing capital markets. Academic research that attempts to incorporate sentiment in pricing models not only as a scalar variable, but as a factor that may not be uniform for all companies, and may vary over time, is fundamental to improve this quite recent field of behavioral asset pricing. (AU)