Abstract

Transaction cost economics (TCE) provides a theory for examining the phenomenon of internal audit. TCE is concerned with economizing or choosing the lowest cost option from buy and make alternatives. With the buy or market alternative, costs are kept low by competition. However, with the make or in-house alternative, costs are controllable because of the ability to direct and monitor production. In this regard, Williamson (1975, 1985) has argued that internal auditors provide managers with useful information for cost economizing. When in-house production is chosen over market, TCE maintains that there are rational reasons for the choice. These reasons are described by Williamson (1985) as dimensions, namely: asset specificity, uncertainty, and frequency (i.e., size). With the existence of these dimensions, information is needed to direct production activities. Accordingly, the usefulness of internal audit operational information for economizing the (dependent variable) is hypothesized to be positively related to the existence of these dimensions. In addition, based on Penno (1990), the usefulness of internal audit findings for economizing is hypothesized to be positively related to the hierarchical level to which the director of internal audit reports. These hypotheses were tested with an experiment administered to 244 senior executives in 29 departments of two large governmental organizations. The response rate was an acceptable 40.6 percent. A split-plot general linear model was used for the statistical analyses. This entailed a full factorial, repeated measures model for the within-subjects variables, asset specificity, uncertainty, and size The hierarchical level to which the director of internal audit reports was analyzed as a between-subjects variable. The hypotheses for the positive relation of asset specificity and size to usefulness were supported at the 0.01 level, while there was some support for the positive relation for the hierarchical level to which the director of internal audit reports. There was no support for a positive relation for uncertainty.