[Untitled]

Much of the agency literature assumes that various monitoring devices are partial substitutes in reducing total agency costs. In particular, internal and external auditing often are characterized as monitoring devices that should be partial substitutes. We argue that reliable evidence of this relation is lacking because prior studies using cross-sectional archival data have not carefully considered the implications of microeconomic theory of substitution for the models estimated. Our analysis leads to a reexamination of the relation using time-series data. We find no evidence that systematic substitution of internal for external auditing (or vice versa) occurred during the period 1989–1993. Further analysis indicates that the relative prices of internal and external auditing inputs did not change during the period. Therefore a necessary condition for substitution to occur did not exist. Although we do not detect substitution with our sample, the analysis and methodology we develop contribute to the literature by enhancing researchers’ understanding of substitution among monitoring methods.