A New Approach Based on Cumulants for Estimating Financial Regression Models with Errors in the Variables: the Fama and French Model Revisited

This paper proposes to revisit both the CAPM and the three-factor model of Fama and French (1993) in presence of errors in the variables. To reduce the bias induced by measurement and specification errors, we transpose to the cost of equity an estimator based on cumulants of order three and four initially developed by Dagenais and Dagenais (1997) and lated generalized to financial models by Racicot (2003). Our results show that our technique has great and significant consequences on the measure of the cost of equity. We obtain ipso facto a new estimator of the Jensen alpha.