Revisiting Conditional Conservatism
Panos N. Patatoukas
University of California, Berkeley - Haas School of Business
Jacob K. Thomas
Yale School of Management
April 9, 2014
We incorporate revisions in expected future earnings to better represent concepts underlying the Basu (1997) discussion of conservatism and to disaggregate Basu’s asymmetric timeliness (AT) coefficient into two components that reflect conservatism by capturing asymmetry between bad and good news relating to: (i) the variance of unexpected earnings, and (ii) the covariance between unexpected earnings and revisions in expected future earnings and discount rates. We find bias in the AT measure as well as its two components, because they comove strongly with placebos that are unrelated to conservatism. Because this bias reflects asymmetries in the conditional distributions of stock prices and returns, future research should deemphasize specifications affected by such asymmetries. We also investigate a debate regarding the source of upward bias that arises when the AT measure is implemented using observed rather than unexpected earnings and returns. Our results support Patatoukas and Thomas (2011) but contradict Ball et al. (2013b).
Number of Pages in PDF File: 49
Keywords: Conditional conservatism, asymmetric timeliness, revisions in expected future earnings, placebo tests
JEL Classification: M41working papers series
Date posted: June 23, 2013 ; Last revised: April 12, 2014
© 2014 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo1 in 0.422 seconds