This research aims to compare the performance of a statistical factor
asset pricing model with the Fama-French-Carhart 4-factor model. We perform
a Principal Component Analysis (PCA) to extract latent risk factors using
data of stocks listed on B3 from 2001 to 2015. We test the abilities of the
two models to explain assets' returns both in the time-series and in the
cross-section dimension. We found that the statistical factor models
generates statistically significant abnormal returns in the time-series
analysis while the 4-factor model does not. In the cross section dimension,
neither model generates significant abnormal returns but they also are not
able to generate positive risk premia. Similar results are found if we
consider different sets of time and assets. Therefore, although the 4-factor
model performs slightly better in the set of tests, neither of the models
can be considered fully adequate to explain expected returns of assets in
the Brazilian stock market.