Testing Reputational Effects in a Natural Experiment : How did Arthur Andersen Clients Choose Their New Auditor?

Reputation is a market mechanism widely believed to effectively resolve problems of asymmetric information. It relies, however, on consumers taking into account any available information on unobserved product characteristics (quality). Exploiting the natural experiment of Arthur Andersen's demise, I investigate to what extent if at all former Andersen clients took into consideration information on unobserved audit quality when they chose their new public accountant. In particular, the effect of auditors' financial restatement history on firms' auditor choice is examined in discrete choice settings. After controlling for auditor and client characteristics and for the endogeneity of audit fees, I find that although financial restatements are thought to be very noisy and often uninformative indicators of quality, former Andersen clients appear to have based their succeeding auditor decisions on them