The Errors in the Variables Problem in the Cross-Section of Expected Stock Returns



    Search for more papers by this author
    • Department of Finance, Rutgers University School of Business. I am grateful for helpful discussions and comments from Steve Brown, Eric Chang, Cheng-few Lee, Edward Nelling, Jay Shanken, William Taylor, and especially Gikas Hardouvelis. I am also particularly grateful to an anonymous referee and René Stulz (the editor) for their help in improving the article. Financial support for this research was provided by the Research and Sponsored Programs at Rutgers University. All errors remain my responsibility.


Recent research has documented the failure of market beta to capture the cross-section of expected returns within the context of a two-pass estimation methodology. However, the two-pass methodology suffers from the errors-in-variables (EIV) problem that could attenuate the apparent significance of market beta. This article provides a new correction for the EIV problem that is robust to conditional heteroscedasticity. After the correction, I find more support for the role of market beta and less support for the role of firm size in explaining the cross-section of expected returns. While the EIV correction leads to a diminished role of firm size, the size variable remains a significant force in explaining the cross-section of expected returns.