The 9/11 attacks in the United States, as well as other attacks in different parts of the world, raise important questions related to the economic impact of terrorism. What are the most effective ways for a country to recover from these economic losses? Who should pay for the costs of future large-scale attacks? To address these two questions, we propose five principles to evaluate alternative programs. We first discuss how a federal insurance program with mandatory coverage and a laissez faire free-market approach for providing private insurance will fare relative to these principles. We conclude that neither solution is likely to be feasible here in the United States given the millions of firms at risk and the current structure of insurance regulation. We then evaluate how well the U.S. Terrorism Risk Insurance Act (TRIA), a public-private program to cover commercial enterprises against foreign terrorism on U.S. soil, meets the five principles. In particular, we show that TRIA has had a positive effect on availability of terrorism coverage and also has significantly contributed to reducing insurance premiums. TRIA is scheduled to terminate at the end of the year, but pending legislation would extend the program for fifteen years after December 31 (HR. 2761). In this paper, we show that such a long-term extension might have important impacts on the market. This could increase the take-up rate, as prices might be even lower than they are today. We show also, however, that if TRIA were extended for a long period of time in its current form, some insurers could "game" the program by collecting ex ante a large amount of premiums for terrorism insurance, while being financially responsible for only a small portion of the claims ex post. The general taxpayer and the general commercial policyholder (whether or not covered against terrorism) would absorb the residual insured losses. This raises major equity issues inherent in the design of the program.