Interdependent Security Particularly Grateful to Richard Zeckhauser for Helpful Discussion and Comments on Earlier Drafts. Useful Insights Were Also Provided

Do firms have adequate incentives to invest in protection against a risk whose magnitude depends in the actions of others? This paper characterizes the Nash equilibria for this type of interaction between agents, which we call the interdependent security (IDS) problem. When agents are identical, there are two Nash equilibria for a wide range of cost and risk parameters-either everyone invests in protection or no one does. In some situations the incentive to invest in protection approaches zero as the number of unprotected agents increases. We develop an IDS model by first focusing on airline security and comparing the structure of this problem with other IDS examples such as computer security, fire protection, vaccinations, protection against bankruptcy, and theft protection. The paper also examines the roles of insurance, liability, fines and subsidies, third party inspections, regulations and coordinating mechanisms for internalizing the negative externalities characteristic of these problems. The concluding section suggests directions for future theoretical and empirical research.