This paper estimates risk reductions for each layer of security designed to prevent commercial passenger airliners from being commandeered by terrorists, kept under control for some time, and then crashed into specific targets. Probabilistic methods are used to characterize the uncertainty of rates of deterrence, detection, and disruption, as well as losses. Since homeland security decisionmakers tend to be risk-averse because of the catastrophic or dire nature of the hazard or event, utility theory and Monte Carlo simulation methods are used to propagate uncertainties in calculations of net present value, expected utility, and probabilities of net benefit. We employ a “break-even” cost-benefit analysis to determine the minimum probability of an otherwise successful attack that is required for the benefit of security measures to equal their cost. In this context, we examine specific policy options: including Improvised Physical Secondary Barriers (IPSBs) in the array of aircraft security measures, including the Federal Air Marshal Service (FAMS), and including them both. Attack probabilities need to exceed 260 percent or 2.6 attacks per year to be 90 percent sure that FAMS is cost-effective, whereas IPSBs have more than 90 percent chance of being cost-effective even if attack probabilities are as low as 6 percent per year. A risk-neutral analysis finds a policy option of adding IPSBs but not FAMS to the other measures to be preferred for all attack probabilities. However, a very risk-averse decisionmaker is 48 percent likely to prefer to retain FAMS even if the attack probability is as low as 1 percent per year—a level of risk aversion exhibited by few, if any, government agencies. Overall, it seems that, even in an analysis that biases the consideration toward the opposite conclusion, far too much may currently be spent on security measures to address the problem of airline hijacking, and many spending reductions could likely be made with little or no consequent reduction of security.