Quantile regression, asset pricing and investment decision

摘要

The present study compares the Fama-French three factor coefficient estimates obtained from both ordinary least squares (OLS) and quantile regression for 25 size-value sorted portfolios of BSE 500. The study, using empirical results, residual graphs and other plots, confirms the inefficiency of OLS in end distribution estimation. Quantile regression reveals that the slope direction for all coefficients of predictor variables is not the same across the quantiles and time. Finally, the study shows, empirically, that quantile regression estimates give a more comprehensive and clearer picture of the varying effect of predictors on response variables to analysts or investors in making investment decisions.