Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Some Additional Evidence on Survival Biases

Some Additional Evidence on Survival Biases I. Introduction A FREQUENT PROBLEM IN the analysis of f m ’ financial time series is the need to impose survivorship criteria and the difficulty of determining their effects. Survivorship criteria usually are imposed in order to obtain sufficient observations for estimation purposes. But realizations of financial time series are unlikely to be independent of whether fmsurvive. Thus, the requirement of a long series for estimation purposes could result in an unrepresentative set of firms’ series, and inferences concerning population characteristics could be biased. In our study [I] of the time series firms’ annual earnings, we required sample members to have twenty consecutive years of earnings data. This required all sample members to have survived for at least twenty years, which would have reduced the likelihood of firms with poor earnings performance entering the sample. An additional and more subtle bias arose because of the criteria imposed by Standard and Poor’s when they constructed the Compustat file, which we used in that study. Their sample was selected essentially on the grounds that firms were of interest to security analysts, who were the likely market for the file, at the time of its construction. The file was then constructed http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Journal of Finance Wiley

Some Additional Evidence on Survival Biases

The Journal of Finance , Volume 34 (1) – Mar 1, 1979

Loading next page...
 
/lp/wiley/some-additional-evidence-on-survival-biases-18nZc1WR85

References (6)

Publisher
Wiley
Copyright
1979 The American Finance Association
ISSN
0022-1082
eISSN
1540-6261
DOI
10.1111/j.1540-6261.1979.tb02080.x
Publisher site
See Article on Publisher Site

Abstract

I. Introduction A FREQUENT PROBLEM IN the analysis of f m ’ financial time series is the need to impose survivorship criteria and the difficulty of determining their effects. Survivorship criteria usually are imposed in order to obtain sufficient observations for estimation purposes. But realizations of financial time series are unlikely to be independent of whether fmsurvive. Thus, the requirement of a long series for estimation purposes could result in an unrepresentative set of firms’ series, and inferences concerning population characteristics could be biased. In our study [I] of the time series firms’ annual earnings, we required sample members to have twenty consecutive years of earnings data. This required all sample members to have survived for at least twenty years, which would have reduced the likelihood of firms with poor earnings performance entering the sample. An additional and more subtle bias arose because of the criteria imposed by Standard and Poor’s when they constructed the Compustat file, which we used in that study. Their sample was selected essentially on the grounds that firms were of interest to security analysts, who were the likely market for the file, at the time of its construction. The file was then constructed

Journal

The Journal of FinanceWiley

Published: Mar 1, 1979

There are no references for this article.