The effect of industry classification on analyst following and the properties of their earnings forecasts

ABSTRACT

Using a comprehensive data set, we compare four broadly available industry classification schemes (Standard Industrial Classification (SIC), North American Industry Classification System (NAICS), Fama–French classification (FF) and Global Industry Classification Standard (GICS)) in their effectiveness to group analysts and their earnings forecast properties. We demonstrate the advantage of the GICS to be consistent across different forecasting properties and across different groups of firms. Our results suggest that GICS should be utilized in research designs, either in the primary analysis or as a necessary corroboration.

KEYWORDS:

• Industry classification
• homogeneity
• analyst forecasts
• GICS

JEL CLASSIFICATION:

• C15
• G1

Acknowledgments

The authors are grateful to R. Zhang and R. Lundholm for helpful comments and suggestions. This work was supported by the Social Sciences and Humanities Research Council of Canada [grant number 31-639958] and the Institute of Chartered Professional Accountants of British Columbia [grant number 21-210069]. Any errors are ours.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Capital market research predominantly utilizes the Standard Industrial Classification (SIC) system in order to be consistent with prior studies and because, prior to 1999, it was the only industry classification system available through major data vendors. However, the reliability of SIC has been a long-standing problem in financial research, as the SIC codes differ across databases and can fail to identify firms with similar operating characteristics (Hrazdil, Trottier, and Zhang 2013).

² For further details on the evolution, availability, limitations and uses of the different industry classifications, see Hrazdil and Scott (2013).

³ The GICS History provides the most comprehensive historical coverage for more than 25 000 active and inactive North American firms going back to June 1985. The results in this note are based on the GICS History as of 30 June 2013.

⁴ Our results are robust to a comparison of finer industry levels (8-digit GICS relative to 3-digit SIC and 4-digit NAICS, as analysed by Hrazdil, Trottier, and Zhang 2014), to inclusion of all analysts’ forecasts prior to earnings announcements, and to partitioning the sample into three equal firm size groups.

Additional information

Funding

This work was supported by the Social Sciences and Humanities Research Council of Canada [grant number 31-639958] and the Institute of Chartered Professional Accountants of British Columbia [grant number 21-210069]. Any errors are ours.