May 15th, 2008

Understanding Unknown Convention Accumulations

Posted at 6:33 PM ET

Emil Metropoulos, Senior Vice President
Contact

The continued rise of the convention industry has created a unique clash exposure for workers compensation (re)insurers. Carriers are increasingly familiarizing themselves with “known” accumulations, the risks associated with the day-to-day workplace, since the terror attacks of September 11, 2001. “Unknown” accumulations, on the other hand, have been more difficult to grasp. When employees across an insurance portfolio gather for conventions, additional catastrophic risk exposures begin to emerge.

Most workers compensation models focus on the known accumulations that could result from such workplace incidents as earthquakes or terrorism. As long as the average employee goes to the office every day for a generally predictable period of time, the modeling process is plausible. A corporate park, for example, may be modeled with the assumption that nearly all employees assigned to a particular building are present at the time of an event. The same thinking is applied to residual effects as accumulations at nearby buildings are impacted. More refined employee accumulation data and modeling is also possible, accounting for shift work, varying occupancy at different business hours and maximum one–time exposures. Even with the best workers compensation data available, though, the conclusions are tenuous in comparison with property lines. In the course of a typical work day (and during a crisis), people move; buildings do not.

Unknown accumulations related to employee movements have been nearly impossible to quantify until recently. However, they are becoming increasingly important to accumulation management strategy. In 2007, more than 5,200 conferences and conventions were held in the United States. Each event can attract tens of thousands of people from several states—and possibly countries—to a space spanning less than a city block. Many convention destinations are known terrorism targets. More than 14 percent of all conventions are hosted in New York and Las Vegas, which are characterized by high terrorism risks. What results is an unusual workers compensation problem, in which a single event could lead to a large number of casualties from disparate locations in a work-related situation.

Modeling innovations have made it possible to measure workers compensation exposures for previously unknown accumulations. Rigorously analyzing quantitative inter-industry clustering based on trade data and convention venue databases reveals peak conference and convention exposures. This can be interconnected with workers compensation data to determine the total exposure for a particular convention. Once identified, sophisticated deterministic loss scenarios including pandemic flu and disease, earthquake or terror attacks can be contemplated and modeled.

Advancements in data management and scenario modeling also have made unknown accumulation modeling more attainable. Information from previously untapped convention databases (including the convention’s catchment area, types of industries attending and the number and seniority of attendees and exhibitors) can be examined alongside workers compensation data to generate more accurate estimates of accumulation risk.

Modelers can use industry and location data to match the industry and catchment area criteria in the convention database to calculate the maximum number of attendees and devise a range of scenarios to gauge insured losses. By showing users the number of attendees drawn from each of the books, as well as the terminals through which they will travel, modelers can identify vulnerabilities by industry and by locale. In the past, carriers had to accept the unknown. Now, some degree of risk management is possible.

To understand the exposure from unknown accumulations, one must consider (and model) the normal flow of convention business. Not everybody stays for an entire conference. Some arrive late; others leave early. Attendees may not stay for the entire day. Larger conventions may occur in several venues. A core group may be at the convention center, but business is likely being conducted in private meetings, at area hotels. And, let’s face it; a few rounds of golf are probably played. Thus, the actual exposure at any given time is likely to be less than the number of registered participants. Through the intersection of convention and risk data analysis, previously unknown scenarios begin to resemble that of a corporate park. Many buildings may be involved, and the population is transient.

Consider a hypothetical terror strike on a convention center. For the sake of simplicity, assume a scenario in which 60 percent of the attendees are in the main hall, where fatalities in the wake of an attack would be most likely. Permanent total and partial injury categories may comprise 20 percent of the injuries. Another 10 percent could be medical-only, and the remaining 10 percent of participants may not have been present at all at the time of the attack. For different days, at various times of day, the assumptions will change. As the situations are examined, the exposure to unknown accumulations becomes more realistic. 

Workers compensation carriers cannot afford to ignore the convention business. Major convention destinations are increasingly attracting more businesses and employees than ever, and their success continues to grow. Venues are growing as well, taking the potential for catastrophe losses even higher. Potential losses from an unknown convention exposure clashing with known workers compensation, casualty and property accumulations near or at a convention venue could constitute a portfolio’s largest probable maximum loss. In addition to paying attention to the exposure, (re)insurers can do something about it. Modeling with both workers compensation and convention databases makes risk mitigation a realistic goal. This field is still new, and such modeling capabilities are not widely available. But, they do exist. Now, risk acceptance is no longer the only alternative.

AddThis Feed Button
Bookmark and Share


Related Posts