Traditional regulation in the food safety domain has been in the form of mandatory, inflexible food safety controls that are applied to firms. There has been a trend away from this regulatory paradigm towards more co-regulation and self-regulation by industry. This paper investigates the potential for systemic failure in the provision of safe food that might arise as a consequence of this new regulatory paradigm. These systemic failures occur owing to the fact that the food safety outcome depends on the behaviour of the three sets of agents (firms, consumers and the regulator). These populations of agents have generally been treated in the literature as homogeneous in terms of their behaviour and strategies. Further, the actions taken by any one agent are assumed to be independent of those taken by others. The institutional economics model that is developed assumes heterogeneity and inter-agent strategic interactions. Given this (more realistic) depiction of behaviour, instances of potential regulatory inefficiencies arise . In particular, the model challenges the trend towards voluntarism and self-regulation.