This study tests whether there is a positive association between a firm's agency costs and its demand for a quality-differentiated audit. Audit firm quality is represented in two ways: a continuous size model in which a direct association is posited between auditor size (measured by clients' sales) and audit quality, and a model in which the Big Eight group of auditors is defined as higher quality suppliers. The tests are supportive of the brand name model of audit quality: agency cost proxies are significant as a group, after controlling for client size and growth, only in the brand name model. The results are also supportive (albeit weakly in some instances) of the following individual agency-related incentives for higher quality audits: monitoring of incentive performance contracts, diffusion of ownership, owner-debtholder conflict, and the subsequent issue of public securities after the auditor change. However, the explanatory power of the models tested is low, after controlling for client size and growth. T HIS paper tests whether the demand for quality-differentiated audits is, in part, associated with variables that proxy for the firm's agency costs (after controlling for client size and growth). Auditing is widely viewed as a means of reducing agency costs [Jensen and Meckling, 1976; Ng, 1978; Simunic and Stein, 1987; Watts, 1977; and Watts and Zimmerman, 1983]. It follows that when agency costs are greater there is increased demand for higher-level audit quality. To test the hypothesis that higherquality audits are demanded as a function of increasing agency costs, a sample of firms is selected that changed auditors. The framework adopted for testing can be described as follows: given that a change in auditor has occurred, is the choice of the new auditor associated with the firm's agency costs? Audit quality is measured in each of two different ways to correspond to the two general theories of audit quality: (1) a continuous auditor size metric based on combined sales of all public companies audited (a proxy for client-specific quasi-rents) and (2) an ordinal categorical variable representing either the brand name Big Eight or nonBig Eight accounting firms. Two repreJere R. Francis is an Associate Professor at University of Iowa, and Earl R. Wilson is an Assistant Professor at University of Missouri-Columbia. Manuscript received June 1986. Revisions received March 1987, September 1987, and April 1988. Accepted May 1988.
Stored only when both papers are in the corpus — 442 citations exist upstream.