Design Issues in Studies on the Effect of Engagement Partner Identification on Audit Quality : Insight from a Natural Experiment

We exploit a 2008 regulatory change in a Canadian province to investigate the association between engagement partner identification (EPI) and audit quality (AQ) and to illustrate the implications of control group design choices on this association. In contrast to prior studies, firms in our treatment and control groups operated in the same institutional and economic environment throughout the period under study, reducing the likelihood of “contamination” of the EPI effect by confounding events and enabling its measurement over a longer period. We find no association between EPI and six AQ proxies when the treatment and control groups are restricted to economically similar firms, in contrast with the positive association documented in tests with inferior research designs. We conclude that EPI had no impact on AQ in our Canadian setting and we recommend that researchers proceed with caution when the treatment and control groups operate in different institutional or economic environments