Predicting the future, in cross-section
Year: 2016
Type: dissertation
Abstract: Many of the studies involving return predictability include only a few predictive variables that can, with relative accuracy, predict future returns. This dissertation uses 12 firm specific characteristics to study the out-of-sample predictive power of return forecasts derived from cross-sectional regressions. We group these characteristics into 5 models in order to test which combination of characteristics is the best at predicting actual returns. Empirically, the slopes derived from the regressions present strong predictive power, especially using 10-year rolling estimates with models that include either 9, 10 or 12 variables. We also test the feasibility of a trading strategy based on expected returns, which turned out to be very profitable, displaying Sharpe ratio values as high as 1.91 annualized, beating the market in all the sample period.
Author: Ferreira, Jácome de Bruges Toledo Mendes
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W2346699928
