A non-parametric approach to evaluating catastrophe risk and decisions: ..

...Financial and Infrastructure SystemsSPEAKER: Dr. Craig Taylor,Research Professor, USC, Sonny Astani Department of Civil and Environmental Engineering and Sr. V.P., Baseline Management Company, Inc. (with help from many)Abstract:
Catastrophes raise many issues for financial and infrastructure systems. These pertain to the use of insurance, bonds, swaps, design, retrofit, upgrades and many other activities that require complex evaluations. A great deal of past and current work has emphasized the use of parametric formulations of catastrophe loss distributions as means to help on major catastrophe risk decisions. This presentation goes in an opposite direction—even opposite to my own research until very recently.
With ongoing research, I currently maintain that enormous advances in computer technology (speed and storage) illuminate how these parametric fits tend to produce dramatic sub-optimization in catastrophe decision-making. Moreover, these advances further help to show that efficiency gained through these fits is minor.
Other themes proposed in this presentation are that:
• Parametric modeling can be very helpful on less major projects and especially on the sub-models developed in these complex evaluations
• Averages are very important, but so are extremes (potentially impacted by tipping points, yield points, deductibles, attachment points, limits of liability)
• Focus on efficiency should be on rendering systems evaluations (e.g., transportation systems given multiple bridge collapses) accurate yet faster
• Of necessity probabilistic, criteria for finance should incorporate gains as well as losses
• Use of equi-probable estimates such as for ROI (return on investment) in finance or for total costs in engineering greatly facilitates estimation and understanding of financial criteria
• For the extreme distributions involved with catastrophes, variance reduction techniques have so far proven to be far less important than assumed; advances in computer technology (speed, storage, pre-processing) have had vastly greater impacts.
• Complex catastrophe (or shock-based) evaluations should take advantage of these developments so that engineers, economists and others can focus on sub-models and assumptions.

Location: Kaprielian Hall (KAP) - 209

Audiences: Everyone Is Invited

Contact: Evangeline Reyes