Measurement error in the returns/earnings association: Diagnosis and remedies

1996 1996

Other formats: Order a copy

Abstract (summary)

A primary focus of empirical earnings research is whether accounting earnings contain meaningful and timely information that can be used by market participants to value securities. Empirically, accounting earnings numbers are considered to be reliable representations of economic performance if abnormal returns and unexpected earnings are associated. A critical assumption made when estimating this association by ordinary least squares (OLS) is that the explanatory variable (unexpected earnings) is independent of the error term. If unexpected earnings is measured with error, this assumption is violated, and OLS will produce biased estimates of the parameters.

It's widely recognized that empirical measures of unexpected earnings contain measurement error, biasing the OLS estimate of the earnings response coefficient (ERC) towards zero. Prior studies have used various techniques to address measurement error in the unexpected earnings proxy. These techniques include: using multiple proxies for unexpected earnings; including lagged security returns; grouping by size of abnormal returns; using reverse regression; and using instrumental variables. No general conclusion has been reached, however, about which technique is the 'best' or whether any of these techniques completely eliminates bias in the ERC. Therefore, in this study I provide a discussion of the conditions under which each error-reduction technique is most effective in reducing measurement error bias in the ERC. In addition, I provide empirical evidence on each technique's ability to reduce coefficient bias. The empirical results show that none of these techniques is useful at reducing measurement error bias in the ERC. The ERC estimate increases by at most 8%.

I also introduce and assess a new technique that yields consistent parameter estimates in the presence of measurement error (Fuller (1987, 1991)). The empirical results indicate that this technique is fairly successful at reducing measurement error bias in the ERC. The ERC estimate increases by as much as 52%. This technique is most successful at eliminating bias in the ERC estimates of larger firms and for firms whose earnings changes are primarily transitory. This technique provides a promising alternative approach to address measurement error bias in the ERC when investigating the information content of earnings.

Indexing (details)

Economic theory
0272: Accounting
0508: Finance
0511: Economic theory
Identifier / keyword
Social sciences; earnings; returns
Measurement error in the returns/earnings association: Diagnosis and remedies
Machuga, Susan Marie
Number of pages
Publication year
Degree date
School code
DAI-A 57/07, Dissertation Abstracts International
Place of publication
Ann Arbor
Country of publication
United States
9780591047592, 0591047594
Elgers, Pieter T.
University of Massachusetts Amherst
University location
United States -- Massachusetts
Source type
Dissertations & Theses
Document type
Dissertation/thesis number
ProQuest document ID
Database copyright ProQuest LLC; ProQuest does not claim copyright in the individual underlying works.
Document URL
Access the complete full text

You can get the full text of this document if it is part of your institution's ProQuest subscription.

Try one of the following:

  • Connect to ProQuest through your library network and search for the document from there.
  • Request the document from your library.
  • Go to the ProQuest login page and enter a ProQuest or My Research username / password.