Adam Rose Coordinator for Economics at The National Center for Risk and Economic Analysis of Terrorism Events (CREATE) and Research Professor in the Price School at USC participated in the recent National Bureau of Economic Research (NBER) Conference on “Insurance Markets and Catastrophic Risk", held May 10-11 in Cambridge, Massachusetts.
The NBER is the leading independent economic research center in the U.S. This is one of many specialty conferences it holds each year, and was organized by CREATE faculty affiliates, Howard Kunreuther and Erwann Michel-Kerjan, both of the Wharton School at the University of Pennsylvania. Professor Rose’s presentation was entitled “Improving Catastrophe Modeling for Business Interruption Insurance Needs," and was co-authored by Charles Huyck of ImageCat, Inc. and its affiliate in the United Kingdom, eCityrisk. The focus of catastrophe insurance, and hence the estimation of potential losses, has been oriented toward property and casualty damages. However, there is a growing interest in being insured against business interruption (BI) losses (essentially lost profits from the curtailment of business activity due to a disaster). This has been inspired to a great extent by the large size of such losses in recent major catastrophes.
Professor Rose recently coordinated eight studies to arrive at a definitive estimate of the business interruption losses from the September 11 terrorist attacks on the World Trade Center. The consensus estimate was approximately $100 billion in business interruption, more than four times the amount of property damage. More recently, estimates of property damage from Hurricane Katrina have concentrated at about $75 billion, but BI estimates now exceed 100 billion and counting (BI is not complete until the economy has recovered or reached a “new normal").
At the same time, BI interruption is difficult to estimate because it does not take place at just one point in time as, does property damage, but continues until recovery is complete. Therefore it is affected by the variability of public policy in terms of reconstruction patterns and outside aid, behavior related to risk perceptions, strategic behavior related to insurance claims, and the effectiveness of numerous ways to promote recovery through resilience. Professor Rose outlined a framework for estimating BI losses, and illustrated how inclusion of various resilience tactics and improved data on economic activity can enhance estimation accuracy. He provided an example based on actual data from an insurance company for a major business enterprise in the southeast U.S. He also provided an overview of progress that can be made in estimating losses from “contingent" business interruption, which refers to more indirect sources of loss from locations that supply critical inputs or accept insured’s products, or manufacturing locations that provide products for delivery to the insured’s customers. This estimation could best be accomplished by the use of computable general equilibrium (CGE) models, an area in which CREATE has done considerable innovative research. CGE models are essentially based on a set of integrated supply chains combined with behavioral reactions to changes in prices and external shocks. They embody many inherent aspects of resilience, and adaptive resilience can be incorporated through parametric changes.
Other speakers at the conference included researchers from the University of California at Berkley, Stanford, Harvard, University of Pennsylvania, Norwegian School of Economics, University of Delhi, World Bank, Federal Reserve Bank of New York, European Central Bank, and International Monetary Fund.
Paper Abstract: While catastrophe (CAT) modeling of property damage is well developed, CAT modeling of business interruption (BI) is still in a relative state of infancy. One reason is the complication of behavioral and recovery policy decisions relating to resilience during the recovery process. Another is the crude nature of functional relationships that translate property damage into BI. This paper proposes a framework for improving the estimation of ordinary and contingent BI. Improved data collection on individual facilities within a company and application of more detailed and realistic resilience adjustments can improve estimation accuracy. We then illustrate the difference this can make in a case study example. We also explain how some macroeconomic modeling approaches are best suited to estimating contingent BI because they can include critical aspects such as supply chain and infrastructure interdependence, as well as the ability to estimate the economic decline following a disaster that affects the demand for goods and services.