Kesete, Yohannes2015-08-202020-05-242015-05-24bibid: 9255495https://hdl.handle.net/1813/40741This dissertation consists of three articles. The first introduces a new modeling framework to help understand and manage primary insurers' roles in catastrophe risk management. The framework includes a new game theoretic optimization model of primary insurer decisions that interacts with a utility-based homeowner decision model, and is integrated with a regional catastrophe loss estimation model. Reinsurer and government roles are represented as bounds on the insurer-insured interactions. The modeling framework can be used to explore two primary questions. First, how should insurers, using a credible assessment of natural disaster risk, optimize their catastrophe risk insurance-policy design, portfolio, and risktransfer decisions within a context defined by homeowners, reinsurers, and government agencies? Second, how do changes in the context affect insurers' ability to operate successfully? Specifically, it provides results that indicate, under equilibrium, the (1) primary insurers' optimal actions and outcomes, (2) homeowners' optimal actions and outcomes, (3) reinsurers' outcomes, and (4) loss distribution for each stakeholder. The second article, using survey data, explores the roles of prior disaster experience and risk perception on flood insurance purchase decision-making. The survey was administered by a computer-assisted telephone interviewing system in the eastern half of North Carolina. A structural equation model was built to understand the direct and indirect effects of different variables on one another and on flood insurance purchase decision-making. The article provides insight on the mediation effect of risk perception, by linking prior disaster experience to the undertaking of protective action. It also discusses the implications of this insight for designing effective risk communication tools, the timing of risk awareness campaigns, and the provision of affordable insurance policies. The third article, using the same survey data from the eastern half of North Carolina, investigates the relationship between self-insurance and market insurance. An ordered logistic model was developed by using revealed preferences about structural retrofit measures and standard homeowners' insurance deductible choices. The article shows that self-insurance and market insurance are substitutes and discusses the implications of this finding in terms of setting appropriate standard homeowners' premium and deductible values.en-USInsuranceRisk perceptionOptimizationdissertation or thesis