PREDICTING STOCK RETURNS THROUGH PAST MOVEMENTS: A MODIFICATION OF GRINBLATT AND MOSKOWITZ MODEL

Abstract

The purpose of this paper is to present the Grinblatt and Moskowitz Model (2004), and make a modification to adapt for an emerging market, in this case to apply in the Sao Paulo Stock Exchange (Bovespa), that presents some specifics characteristics and problems, common in financial models and time series. It was made a modification in the original model and applied to Brazilian Capital Markets. Some interesting results were found: the presence of downside risk and the presence of turbulence during the electoral process in 2002. This modification was significantly because it incorporated the dummy variable to electoral process and took off the benchmark variable, which presented some disturbance when applied to Brazilian database. The modification of the Grinblatt and Moskowitz Model (2004) showed better results than the original one. It suggests that the modification can incorporate some characteristics of emerging markets countries.