Limiting Auditor Liability? - Experimental Evidence on Risk and Ambiguity Attitudes under Real Losses

Abstract

This paper is motivated by the current debate on limiting auditor liability. In a laboratory experiment, the effect of limited versus unlimited liability on behavior under risk and ambiguity is investigated for risks involving small probabilities. The amount of liability is manipulated in such a way that subjects can pay with their show-up fee under limited liability, but they can suffer out-of-pocket losses under unlimited liability. Findings are that both risk aversion and ambiguity aversion are higher under unlimited liability than under limited liability, and these two constructs are correlated under unlimited liability. These findings provide new empirical evidence for the intuition that aversion toward risk and ambiguity plays an especially important role in auditor behavior under unlimited liability. Our experimental results emphasize the importance of considering risk and ambiguity attitudes in economic models on liability, especially when liability regimes are compared that differ in liability limits or ambiguity.