Forecasting risk: The value of long-range forecasting for the insurance industry

Traditionally, the insurance industry has established the probability of a weather event by the use of past weather data (climatology) because forecasts, particularly over the longer term, were not always reliable. Weather forecasts were generally informative over a period of days, not the months between the issuing of a policy and the start of a storm season.

However, recent research suggests that some phenomena can be forecast months in advance more accurately than by using climatology - for example, sea surface temperatures in the Pacific and Atlantic tropical storm numbers. While it is impossible to forecast with certainty, the range of outcomes is becoming more accurate and decision-relevant for the insurance industry.