Resumen
The objective of this study is to answer the criticism to the CAPM based on findings that the market portfolio is far from the efficient frontier. We run a numeric optimization model, based on Brazilian stock market data from 2003 to 2012. For each asset, we obtain adjusted returns and standard deviations such that (i) the efficient frontier intersects with the market portfolio and (ii) the distance between the adjusted parameters and the sample parameters is minimized. We conclude that the adjusted parameters are not significantly different from the sample parameters, in line with the results of Levy and Roll (2010) for the USA stock market. Such results suggest that the imprecisions in the implementation of the CAPM stem mostly from parameter estimation errors and that other explanatory factors for returns may have low relevance. Therefore, our results contradict the above-mentioned criticisms to the CAPM in Brazil.