Heterogeneity in Expertise in a Credence Goods Setting : Evidence from Audit Partners

We examine the heterogeneity of experts in a credence goods setting. In our analytical model, clients are ex ante uncertain about how much effort experts need to provide to solve their problem, which is either simple or difficult. Experts have varying degrees of expertise. Less qualified experts are equally effective at solving simple problems, but they are less effective at solving difficult ones. We show that clients pay a fee premium to more qualified experts, even for simple problems. This premium increases with the ex ante probability that the client has a difficult problem. We empirically test the model predictions in the context of partner industry expertise for the U.S. operations of the Big 4 audit firms. We find, consistent with the model, a positive association between partner industry specialization and audit fees, even for simple audits, and a negative association between partner specialization and the client's probability of restatements only for difficult audits. Further, the industry specialization premium is higher in industries with higher proportions of difficult audits. Consistent with credence good agency issues, the specialization premium for simple audits is mitigated when information asymmetry between client and auditor is lower