An Experimental Investigation of Auditors' Liability: Implications for Social Welfare and Exploration of Deviations from Theoretical Predictions
Posted: 8 Feb 2000
Date Written: February 3, 2000
This paper reports the results of an experiment designed to test the Schwartz (1997) model which makes predictions about audit quality, investments, and social welfare. The setting we investigate is a two-person strategic auditing game in which auditor-subjects select a level of audit quality and investor-subjects select an investment level. The combination of audit quality and investment determined social welfare. We investigate four legal regimes, each consists of a liability rule (strict or negligence) and a damage measure (out-of-pocket or independent-of-investment). We predict, based on Schwartz (1997), that three of the regimes would induce audit quality and investment choices that maximize social welfare, while the fourth regime that consists of a strict liability rule and an out-of-pocket damage measure (Strict/OOP) results in lower social welfare because of excessive investment.
We do not find significant differences in the social welfare across the four legal regimes, which is contrary to the theoretical predictions. However, we find that in the Strict/OOP regime payoffs to investors are significantly higher and payoffs to auditors are significantly lower than in the other three regimes. Thus, if the objective of policy makers is to maximize efficiency, all four regimes are equally desirable. But, if wealth distribution and equity issues are important to policy makers, the Strict/OOP regime may be less desirable, unless auditors can offset the increased liability by sufficient increases in audit fees. We also explore whether subjects' deviations from the predicted actions can be explained by the subjects' limited ability in anticipating their counterparts' actions. We find that the auditors' low payoffs in the Strict/OOP regime cannot be justified with this explanation because auditors' anticipated investments most accurately in that regime. Thus, the deviations of subjects' actions from the predictions are due to choosing actions that do not maximize expected wealth, in spite of correctly anticipating investment levels and potential damages.
JEL Classification: M94, C91, K22
Suggested Citation: Suggested Citation