We study how beliefs affect individuals' willingness to undertake prevention expenditure through a two-type, N-person public good game and test several results empirically. We show analytically that pessimistic agents will invest more in prevention expenditure than optimists. We should how pessimistic beliefs lead to a `double deprivation' and discuss potential issues and remedies. The more optimistic the society the lower will be its total green expenditure. We also demonstrate how small differences in beliefs may induce substantial differences in type-related prevention expenditure. The more atomistic agents are the less they will contribute to the public good. We then use a large international survey to study determinants of prevention expenditure. We proxy beliefs through three variables, namely science optimism, eco optimism and feelings of atomism. For each variable wefind, as predicted by the theoretical model, a significant relationship with the willingness to undertake prevention expenditure. However, environmental education shapes these relationships. While environmental education does not affect the relationship between eco optimism and prevention expenditure, it leads to a stronger relationship between both science optimists, and those who feel atomistic, and prevention expenditure. Finally, we develop a dynamic game with endogenous beliefs based on the static model and discuss the main differences in the optimal choices of the agents. Wefind that pessimistic agents have a higher prevention expenditure compared to the static case since they take the endogenous feedback of the prevention expenditure on their beliefs into account
