Discussion of Lawsuits against Auditors

Abstract

Lys and Watts (hereafter LW) report empirical evidence on selected audit and client characteristics associated with litigation against auditors over allegedly misleading financial reports during the period 1955-94. LW use a paired-sample design which matches each of the 153 lawsuit firms with a nonlawsuit firm of similar size, industry, and delisting status; their univariate and multivariate tests compare characteristics of the resulting lawsuit and nonlawsuit samples. Their analysis extends Stice's [1991] work in four ways: (1) they use a larger sample of lawsuits against auditors which encompasses recent time periods; (2) they investigate a larger set of variables; (3) they measure the variables in both the median wrongdoing year and the filing year (Stice compared characteristics in the year prior to alleged wrongdoing);1 and (4) they examine the relative importance of client and audit characteristics across subperiods believed to be characterized by different levels of auditor liability or exposure. Because LW's results are sensitive to some features of the test design (e.g., univariate versus multivariate tests and when the variables are measured), I have chosen to focus my discussion on the results of their multivariate analysis performed on variables measured in the median wrongdoing year. In terms of audit characteristics, these results show that the incidence of lawsuits naming the auditor as a defendant is