FIN-02-019

Predictive Regressions: A Reduced-Bias Estimation Method

June 2003

Yakov Amihud and Clifford M. Hurvich

ABSTRACT

We propose a direct and convenient reduced-bias estimator of predictive regression coef-ficients, assuming that the regressors are Gaussian first-order autoregressive with errors that are correlated with the error series of the dependent variable. For the single-regressor model, Stambaugh (1999) shows that the ordinary least squares estimator of the predic-tive regression coefficient is biased in small samples. Our estimation method employs an augmented regression which uses a proxy for the errors in the autoregressive model. We also develop a heuristic estimator of the standard error of the estimated predictive coefficient which performs well in simulations, and show that the estimated coefficient of the errors and its squared standard error are unbiased. We analyze the case of multiple predictors that are first-order autoregressive and derive bias expressions for both the or-dinary least squares and our reduced-bias estimated coefficients. The effectiveness of our estimation method is demonstrated by simulations.

Keywords: Stock Returns; Dividend Yields; Autoregressive Models.

Yakov Amihud
Institution: Stern School of Business, New York University, 44th West 4th Street, New York, NY 10012
Telephone: (212) 998-0720
Fax: (212) 995-4220
Homepage: http://www.stern.nyu.edu/~yamihud/
Email: yamihud@stern.nyu.edu

Clifford M. Hurvich
Institution: Leonard N. Stern School of Business, New York University
Telephone: 212-998-0449
Fax: 212-995-4003
Email: churvich@stern.nyu.edu
Home Page: http://www.stern.nyu.edu/~churvich/

To download a copy of this paper click here
To request a copy of this paper click here

The Department of Finance Working Paper Series has been generously sponsored by