Campbell R. Harvey
Duke University, Durham, NC 27708, USA
National Bureau of Economic Research, Cambridge, MA 02138, USA
Chris M. Kirby
University of Maryland, College Park, MD 20742, USA
A number of well-known asset pricing models imply that the expected return on an asset can be written as a linear function of one or more beta coefficients that measure the asset's sensitivity to sources of undiversifiable risk. This paper provides an overview of the econometric evaluation of such models using the method of instrumental variables. We present numerous examples that cover both single-beta and multi-beta models. These examples are designed to illustrate the various options available to researchers for estimating and testing beta pricing models. We also examine the implications of a variety of different assumptions concerning the time-series behavior of conditional betas, covariances, reward-to-risk ratios. The techniques discussed in this paper have applications in other areas of asset pricing as well.