Resumen
This work investigates the ability of the conditional CAPM to explain anomalous returns related to momentum, size and book-to-market effects using Lewellen and Nagel?s (2006) methodology in the Brazilian stock market. To this end we studied a sample of Bovespa?s stocks in a monthly basis from July 1995 to June 2008. The results indicate that only the book-to-market effect presents statistical significance. The conditional model, tested from time series of 12 months, also showed no significant gains in relation to the unconditional form. However, there are evidences that betas do vary over time, suggesting that the sample size on beta calculations may influence portfolio choices, i.e., the evidence of variation in betas over time means that analysis based on the CAPM should be cautious when using unconditional models.