【文章內容簡介】 on needs to pay more attention to officelevel auditor behavior as the unit of analysis and to the size of a local engagement office. Second, this paper is the first to consider office size as a critical factor in audit pricing. Given that no previous research has examined whether audit fees are influenced by the size of a local office, our evidence helps us better understand the nature of auditorclient relationships in the context of audit pricing. Finally, the findings in this study provide both regulators and practitioners with useful insights into what determines audit quality and thus audit fees. Our results suggest that regulators would have a better assessment of audit quality if they shift the level of quality parison to small versus large auditors at the office level, and away from Big 4 versus nonBig 4 auditors at the national level. Economic theory on quality premiums claims that producing goods and services of a uniform quality for various markets and consumers over time is crucial for maintaining quality premiums(., Klein and Leffler 1981。 Kreps and Wilson 1982 。 Shapiro 1983). Similarly, our evidence suggests that large, Big 4 auditors should take care to maintain a similar level of audit quality across offices of different sizes because a systematically poorquality audit service performed by a local office could potentially cause damage to the reputation for the entire audit firm. HYPOTHESIS DEVELOPMENT Office Size and Audit Quality A growing body of audit research emphasizes the importance of analyzing the behavior of auditors in citybased, local engagement offices. However, none of these studies has paid attention to the size of a local office in the context of audit quality. Why does the office size matter in audit quality over and beyond two wellknown audit firm characteristics, ., audit firm size or brand name(Big 4 versus nonBig 4) and industry expertise? In DeAngelo’s(1981b) framework, an auditor’s incentive to promise audit quality with respect to a particular client depends on the economic importance of the client relative to the auditor’s client portfolio. Her analysis indicates that large auditors are likely to provide higherquality audit services to a particular client than small auditors because an auditor’s economic dependence on that client is negligible for large auditors, and large auditors have more to lose (., bear higher reputation loss ) in case of audit failures, pared with small auditors. DeAngelo’s (1981b) theory can also be applied to the analysis of audit quality differentiation between large versus small offices, because a local engagement office can be viewed as a semiautonomous unit in terms of its audit decisions, client base, revenue sources, and other factors (Francis 2021。 Francis et al. 2021。 Wallman 1996) . Large local offices are less likely to depend on a particular client than small local offices because the former have deeper officelevel clienteles and thus are less economically dependent on a particular client. In other words, large offices are less likely to acquiesce to client pressure for substandard reporting than small offices. Further, local offices, whether small or large, may not bear the full amount of reputation losses associated with an audit failure because a substantial portion of the reputation losses are likely borne by the nationallevel audit firm itself. While the reputation losses in the event of audit failures are likely to be greater for large audit firms (DeAngelo 1981b ) , the losses are not necessarily greater for large local offices than for small local offices, because these costs are more firmwide in nature rather than officespecific. This means that local offices may be more concerned with the economic importance of a particular client than a potential litigation risk from audit failures, in particular, when the offices are small in size. The above arguments lead us to predict that large local offices with relatively deep local clienteles are less likely to promise audit quality with respect to a particular client, and thus that they are likely to provide higherquality audit services, ceteris paribus, pared with small local offices with relatively thin local clienteles. In such a case, one would observe a positive association between office size and audit quality. We call this prediction the economic dependence perspective . SUMMARY AND CONCLUDING REMARKS While previo