April 1, 2001 by Glenn McGillivray, assistant vice president & head of corporate
In many parts of North America spring marks the rise of a season many fear. It’s the time of year when tornadoes most often rear their ugly heads. Unlike major earthquakes and hurricanes, it must be noted that tornadoes can and do occur in most every U.S. state and Canadian province, in any month of the year, at any time of the day. What is more, they can arrive as single vortexes or they can come in legions, wiping entire towns off the face of the map.
Witness Monday, May 2, 1999, when a total of 76 tornadoes tore through 18 U.S. states, including Oklahoma and Kansas, leaving at least 44 people dead. Incredibly, the twisters were all spurred from a single storm system. When speaking of tornadoes in Canada, two infamous events often come up in the discussion. First is the Edmonton tornado. On Friday July 1, 1987 a tornado cut through Edmonton, Alberta, from south to north, ripping a swath 37 kilometers in length. Immediately to the west, an area of 125 square kilometers (almost 50 square miles) was struck by tennis ball-sized hail. As the name implies, the Edmonton tornado event consisted of the touchdown of a single vortex.
The second event is the May 31, 1985 Barrie tornado. But unlike the name implies, 14 separate vortexes touched down that day with five of them grouped by the news media as collectively comprising the Barrie “tornado.” Twelve people were killed and scores of others were injured as the storms moved across the southern part of the province. More than 300 houses were destroyed and insured property damage totalled just under $84 million.
Losses likely to rise
Tornado losses of a magnitude similar to those of the May 1999 Oklahoma, or the Barrie event are likely to occur more often as the wealth, insurable equity, and in some instances, the population situated in tornado prone areas increase. Previous industry losses exceeding US$1.5 billion have only been associated with the more severe cat peril events such as hurricanes and earthquakes, but are now expected to occur more often with severe tornadic events. As such, meteorologists, federal and local authorities, and the general public as well as the insurance industry are becoming increasingly concerned with the potential catastrophic severity of tornadoes and hailstorms.
Historically, tornado and hail losses have been more of a frequency problem then one of individual severity. According to NOAA’s Storm Prediction Center 1,351 tornadoes occurred across the United States in 1999. This annual total is eclipsed only by the 1,424 recorded in 1998. The frequency is clearly there. From a severity standpoint, the total estimated insurable losses due to tornado and hail events over the last two years averaged well over US$4 billion per annum.
These past two years are not exceptions. Over the past 25 years the total property catastrophe losses due to tornadoes and hailstorms (US$42 billion) is greater then the catastrophe losses associated with either hurricanes (US$34 billion) or earthquakes (US$17 billion).
The relatively high loss frequency and individual event severity potential of US$2 billion is a cause for concern to many insurers who are financially unprepared to meet the increased loss potential. A review of some of the more costly regional events and the total aggregated activity within those regions during the 1997-1999 period illustrate the extent of recent losses.
During the first part of the year in 1998 a series of tornadoes ripped through central Florida leaving 1,500 homes uninhabitable. Losses in excess of US$60 million were reported from this one event. The total amount of loss due to tornado activity throughout Florida over an extended period of time (fall of 1997 through the winter of 1998) exceeded US$500 million.
In late April of 1998 parts of Alabama were hit by tornadoes. A short while later tornado activity broke out in parts of Arkansas and Tennessee. Total damages in the Southeast during the winter-spring period of that year rose to US$1 billion.
In January 1999 a series of outbreaks running from Texas to Tennessee caused over US$1 billion in damages. Later on in that same year a storm system produced a number of tornadoes across Kansas, Oklahoma, Arkansas, and Tennessee with as many as 10 reaching F4-5 strength. Losses exceeded US$1.5 billion as over 10,000 homes and businesses were destroyed.
Balance sheet impact
In each of the last two years total losses due to tornadic activity significantly impacted the balance sheets of some insurers. A number of factors including the diversity of the portfolio, financial strength, reinsurance protection and risk strategies determined the extent that these losses had on an insurer’s results.
Insurers ranging from larger national to smaller regional carriers are equally concerned about the effects that these storms have on underwriting results. Many insurers have insufficient financial protection against events of high frequency and low to mid-range severity. This inefficiency in their protection causes them to absorb the majority of tornado and hail losses on a net basis.
Cat XOL is the traditional risk transfer mechanism used by most insurers to protect their net position against natural catastrophes. These covers are designed to safeguard against the larger severity type of catastrophic losses. The retention level and amount of protection purchased are highly individualized under these covers and are typically designed as a function of risk appetite, cost of cat capacity, financial strength, and the natural hazards threatening the portfolio.
Each insurer is different. Certain regional companies whose portfolios are only exposed to tornadoes purchase coverage at lower deductibles to address the frequency aspects of the tornado problem. However many other regional companies as well as most national companies with exposure to more than one hazard design their coverage to the less frequent but higher severity hazards of hurricane and earthquake. As a whole, according to a Swiss Re sigma study (issue 07/97, Too little reinsurance of natural disasters in many markets) on average the industry does not activate most of its CAT XOL protection until the loss level approaches US$6 billion.
Generally, cat deductibles exceed the losses produced by tornado and hail events. Hence most of the losses associated with a tornado event fall into the gap between coverage and are absorbed net by the insurer.
The paths associated with a tornado are narrow and random. As such the damage caused by most tornadoes is not as broadly distributed within the industry as other natural catastrophe events. Accordingly the losses to insurers will vary from insurer to insurer based on the randomness of the tornado’s path and the exposure distributions of the insurer’s portfolio.
Filling the gap
The randomness and potential net factors add volatility to an insurer’s results despite the relatively consistent loss levels experienced over the last 25 years. To many insurers the annual net aggregated losses caused by tornadoes could possibly exceed those caused by a major hurricane or earthquake in any given year. Insurers have looked for additional risk transfer products to fill in the void between cat covers and their net retention. Certain products such as Swiss Re’s Frequency Protection Cover (FPC) are specifically designed to smooth out the volatility caused by the lower severity / high frequency events by filling the gap between deductible and the net retention of traditional covers.
The advent of innovative products such as the FPC offers the insurer a broader array of risk transfer mechanisms to level portfolio risk volatility. Proper use of this and other traditional and non-traditional covers will help lessen the increasing potential impact associated with larger losses due to tornado and hail activity.
www.canadianunderwriter.ca CANADIAN UNDERWRITER / APRIL 2001