Liability Rules and Prevention for Risky and Ambiguous Activities: Theory and Experimental Evidence

Abstract

We develop a formal model to compare the incentive effect of strict liability rules in a risky and ambiguous environment. The firm's business activity entails a risk of technological disaster, which likelihood is a decreasing function of prevention. To assess our theoretical predictions, we design a lab experiment by varying the legal regime and the environment. Consonant with our theory, we obtain that limited liability stimulates subjects' investment in prevention in comparison to unlimited liability when only risk matters. Besides, subjects opt for more prevention under the regime of unlimited liability when shifting from risk to ambiguity. Nevertheless, this last theoretical finding is not observed with limited liability. Empirical results from panel fixed effect regressions confirm the treatment effects. We also collect some additional characteristics about the subjects' social preferences during the experiment.