An objective of the project is to re-evaluate pertinent financial risk and decision procedures for assessing seismic risk reduction measures.  Previous port seismic risk and decision procedures have benefited from the application of the mean-variance criterion (MVC).  Current work is exploring whether or not stochastic dominance procedures are superior to the MVC approaches in these decisions. A simplified "quantile" method has been proposed for evaluating diverse mitigation investments in terms of stochastic dominance methods.  This method attempts to render stochastic dominance methods more calculable given current capabilities in electronic technology.  This method considers a risk-averse as well as a risk-taking re-interpretation of these and other financial methods in terms of total costs. Issues arising from the use of MVC and stochastic dominance approaches arise from the left-tail of total cost distributions, or, equivalently, estimates at the lowest quantiles.  For the original interpretation, these problems arise because there may be exposure periods during which there are mitigation costs but no gains.  For the re-interpretation stressing a risk-averse approach, extreme risk aversion conditions the application of stochastic dominance methods.  These limitations are, though, by no means so severe as for other financial decision methods.

Previously used financial methods for evaluating catastrophe risk-reduction methods for port and other infrastructure systems have been shown to have limitations.  One promising method, second-order stochastic dominance (SSD) deserves exploration because it purports to take into account the entire distribution of costs and gains.  One previously stated limitation of this approach is that it requires more calculations than say benefit-cost, least total mean cost, or mean-variance criterion (MVC) approaches.  This limitation is overcome in this study through the use of random walks, in the first place, and the use of a simplified numerical integration procedure, in the second place, and new electronic technologies, in the third place.  However, these simplifications show how the SSD criterion may encounter limitations when faced with catastrophes having low probabilities of occurrence, so low that during a specific exposure period no catastrophe damage may result.  These limitations typically only involve adding alternatives to the efficient set that can be removed on judgment by the decision-maker.  There are no doubt many decisions involving tradeoffs between the least total mean cost and extremely rare events that may pose challenges to the application of these principles.  Still, these methods permit clarification of the nature of these challenging decisions.

Return to Research Activities