With the hurricane season officially starting on June 1, residents of the Gulf and Atlantic Coasts are no doubt listening carefully to the many forecasts that are being made regarding the number of “named storms” that are likely to occur this year.  Insurance companies and investors may also look to such forecasts to anticipate damages that may take place during the storm season.  The Wall Street Journal published an interesting article yesterday regarding the accuracy of early season predictions.  For the most part, such forecasts are not reliable indicators of the severity of the storm season.

Obstacles to Accurate Predictions

The Wall Street Journal article indicated that a number of factors are to blame for poor accuracy of early forecasts.  One of the main difficulties is that seasonal weather patterns that are not well known until late spring can heavily influence summer storm activity:

… it’s hard to even detect these weather patterns far in advance — even giant patterns that determine the intensity of a season. El Niño, or warming of Pacific Ocean waters, tends to suppress hurricanes; La Niña, unusually cold Pacific waters, tends to increase storm activity. Yet neither of these seasonal effects can be predicted with much reliability before the late spring. “Until you really get into the spring and the weather patterns start to set up, it’s really hard to get any kind of decent forecast as to what’s going to go on in the summer and fall,” says Chuck Watson, who works on forecasts of damage from hurricanes. Anytime before spring, “You might as well throw a dart.”

If early season predictions are no better than “throwing a dart”, how about predictions that are made once the season begins?  The record here appears to be somewhat better, presumably because the weather patterns that can influence summer storm activity are more firmly established by early June:

Some forecasters update their predictions once the season has begun. And those forecasts do well. But none of the major forecasts that come out before June has improved significantly on a simple prediction scheme that calls for the same number of named storms and hurricanes as the average of the five prior years. And some do much worse.

Named Storm Count Does Not Predict Economic Damages

While most of the forecasts predict the number of named storms for the upcoming season, this measure does not necessarily have much to do with the total economic damage resulting from the storm season.  This is due to two main factors.  First, storm intensity can vary widely within a storm season.  As we have seen in recent years, a Category 4 or 5 Hurricane can result in an order of magnitude more damage than a Category 1 or 2 storm.  Second, the location of the storm’s landfall has a major impact on economic damages with the worst case scenarios coming when storms hit a major population center.  Hurricane Andrew in 1992 and Hurricane Katrina in 2005 are the most notorious examples:

The number of named tropical storms and hurricanes can have little to do with the damage they create: Hurricane Andrew struck in 1992, a year of below-average storm counts. “The total number of storms is a red herring,” says Joe Bastardi, chief long-range and hurricane forecaster for AccuWeather.com. “It’s a joke.” Many forecasts include more useful measures such as the number of storms that hit land or the accumulated cyclone energy, which quantifies total storm intensity. But news reports often focus on the more-accessible predictions of storm counts.

Insurance Industry Implications

While it is natural to seek out predictions regarding the upcoming storm season, it seems doubtful that the various forecasts that are made each year should be taken too seriously, at least not until the storm season begins.  From the perspective of insurance underwriters, I doubt that these predictions make much of an impact when setting premiums or assessing overall levels of exposure.

The fact is that the industry is inevitably going to experience a year of devastating losses every so often followed by years of more benign conditions.  Underwriting discipline is needed during the benign periods to make sure that premiums are adequate to cover losses and provide some level of underwriting profit over long periods of time while knowing that losses will occur during particularly active storm seasons.