Newsletter Sign-up

This was going to be my year, according to the forecasters. Every sixth year, "Erin" makes it on the list of Atlantic storm names. And being fifth in the pecking order with an above-average season widely predicted, there was a fair chance that my dream, however morbid, of seeing my name in headlines would come true.

But I, like the rest of the named Atlantic storms in 2007, fizzled and fell flat in a matter of hours like a two-liter bottle of soda left uncapped at a kid's birthday party. Sure, Texas got a bit drenched, but what else is new?

The forecasters had predicted an above-average season. In May, NOAA predicted between 13 and 17 named storms and between 7 and 10 hurricanes, while the Colorado State geeks called for 17 named storms and 9 hurricanes and TSR predicted 16 and 9 respectively.

It's time that these forecasters find a more useful purpose in life. Like start an ice sculpture enterprise in California wildfire country. Or organize a gambling ring for Florida elementary-school kids to place bets on school days canceled for bad weather. Either way, risk managers will benefit from the absence of hyped-up hurricane predictions. More specifically, insureds will breathe easier. No forecasts means no excuse to inflate property-CAT rates.

That is the sad fact of these less-than-accurate forecasts. After the dud that was the 2006 hurricane season, insurers had more dough than Papa John. You'd hope forecasters would sharpen their tools a bit and if not hit the nail on the head--after all, no one can expect that amount of accuracy--at least land in the vicinity of the two-by-four.

Inaccurate forecasts themselves are not harmful. Conditions of La Niña, sea surface temperatures and wind shear are complex climate factors that are not definitive indicators of increased hurricane activity. But forecasters use them to formulate a best guess. Because this guesswork is being used as a tool to price property insurance and catastrophe premiums, hurricane forecasts should be abolished. Insurers shouldn't be packaging peace of mind, affixed with an inflated price tag, on little more than a hunch.

What's laughable is the looseness now associated with "forecasting." Some of these forecasting groups publish their first reports in December of the previous year, and update them on a regular basis through the actual hurricane season. It's coming closer and closer to telling people what to expect as it's happening. The "cast" without the "fore."

More disturbing yet is that these "updates"--which some might classify as cheating--still fall short of the actual outcome. This year in September, Colorado State's folks barely tweaked their April numbers. The spring forecast of 17 named storms, 9 hurricanes and 5 intense hurricanes was lowered in September to 15, 7 and 4.

Homeowners and risk managers in coastal areas know the risk. They don't need forecasts every season that will encourage either a false sense of security or pointless panic. Wouldn't a complete lack of expectations about a hurricane season be more useful to the insured? The message would be one of a constant state of guardedness. Not blissful unawareness, nor irrational anxiety. Just caution. Insurers would get a similar message for pricing their coverage: year after year just expect an outcome somewhere between a 2005 season and a 2006 season.

If you can't accurately predict a Raging Rita or Effeminate Erin--or a total nonseason--get out of the research lab.

ERIN GAZICA is associate editor of Risk & Insurance®.

(Read Matthew Brodsky?s Point, "Hurricane Forecasts: A Useful Tool.")

December 1, 2007

Copyright 2007© LRP Publications