Measuring Reporting Conservatism using the Dichev-Tang (2008) Model

This paper provides a critical evaluation of an alternative measure of reporting conservatism introduced by Dichev and Tang (2008). Although there is substantial interest in research on accounting conservatism, there is no consensus among researchers on the most appropriate measure of conservatism i...

Full description

Saved in:
Bibliographic Details
Main Author: LEE, Jimmy Kiat Bee
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2011
Subjects:
Online Access:https://ink.library.smu.edu.sg/soa_research/1058
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2100319
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
Description
Summary:This paper provides a critical evaluation of an alternative measure of reporting conservatism introduced by Dichev and Tang (2008). Although there is substantial interest in research on accounting conservatism, there is no consensus among researchers on the most appropriate measure of conservatism in empirical studies. Dichev and Tang (2008) introduce a new measure of conservatism, which they believe to be a 'natural and practical measure of conservatism (p. 1441).' However, the econometric properties of this measure have not been fully evaluated, and previous studies have not provided evidence of this measure’s construct validity. Based on a parsimonious model of conservatism, I find that the Dichev and Tang (2008) measure is increasing in the conservatism parameter. However, although this measure produces well-specified test statistics that generate Type 1 errors according to researchers’ specifications, it generates tests of low power that lead to relatively high Type 2 errors. Next, I use actual data to provide evidence consistent with the construct validity of this measure. Finally, I suggest an alternative measure by using the reverse regression specification of the Dichev and Tang (2008) model. Results from simulations suggest that this alternative measure is feasible and is slightly superior to the Dichev and Tang (2008) measure in terms of test power.