Interdependent security experiments

This paper analyzes the behavior of subjects in interdependent security experiments which exhibit strategic complementarity. In these experiments, subjects decide whether to pay to mitigate the risk of a loss, but the exact risk depends on the actions of other subjects. Two ranked equilibria exist, and the efficient equilibrium is for all agents to pay for the mitigation. Subjects in the interdependent security experiments rarely coordinate on the efficient equilibrium. Coordination is slightly more common in similar coordination games without the risk mitigation decision. The experiments also compare the effectiveness of two policies at inducing higher levels of mitigation.