This paper explores the issue of whether strict liability imposed on polluters has served to reduce uncontrolled releases of toxics into the environment. To answer this question, we exploit the variation in state hazardous waste site laws across states and over time. We use data on accidents and spills involving hazardous substances and fit regressions relating the frequency of spills of selected chemicals used in manufacturing to the type of liability in force in a state and state manufacturing activity. Results vary with the chemical being analysed. For some chemicals, the presence of strict liability does not provide any additional explanatory power for the number of spills beyond what is achieved by the number of establishments and the sectoral composition of manufacturing. For other families of chemicals, we find that spills are more numerous in states that impose strict liability. Further investigation suggests that (i) for some of these chemicals, this could be due to unobserved state characteristics influencing spills, which may have acted to reduce the incentives of liability, and that (ii) small firms are responsible for a disproportionate number of spills, regardless of the liability structure. An alternative explanation, supported by the results of a separate regression for two liability regimes, is that only under strict liability are small firms responsible for a disproportionate number of spills.