Optimal Cross-Sectional Regression
61 Pages Posted: 28 Jan 2021 Last revised: 27 Feb 2023
Date Written: September 30, 2021
Abstract
Errors-in-variables (EIV) biases plague asset pricing tests. We offer a new perspective on ad-
dressing the EIV issue: instead of viewing EIV biases as estimation errors that potentially
contaminate next-stage risk premium estimates, we consider them to be return innovations
that follow a particular correlation structure. We factor this structure into our test design,
yielding a new regression model that generates the most accurate risk premium estimates. We
demonstrate the theoretical appeal as well as the empirical relevance of our new estimator.
Keywords: Beta uncertainty, Efficient esetimation, Factor models, Fama-MacBeth, GMM, Idiosyncratic risk, Systematic risk, Two-pass regression, Errors-in-variables
JEL Classification: C14, C22, G12
Suggested Citation: Suggested Citation