A new hail risk modelling system is available for Canadian crop insurers, but providers are divided over its value.
AIR Worldwide recently developed the industry’s first probabilistic computer model designed to better assess hail risk in Canada.
Ken Doleman, chief executive officer at Palliser Insurance, said the additional data from the AIR model will provide a more complete picture.
“It gives us a better forward look as opposed to just simply a rear-mirror look,” he said.
He said the AIR computer model looks at 10,000 years of weather and climate history and uses calculations to look at the probability and timing of storms.
That information is overlaid with Palliser’s own database, which Doleman described as likely the most extensive held by an individual organization on hail loss insurance in Western Canada.
However, Saskatchewan Municipal Hail Insurance CEO Rodney Schoettler is less convinced. The main feature of the AIR probabilistic model lies in its ability to project possible future risks, but Schoettler said that has little value in his current business model.
“We don’t really honestly care what the future will bring,” he said.
“If a model says two years from now it’s going to hail more so your risk is going to go up, we’re not going to respond today anyway. We’re still going to write it based on our experience.”
He said SMHI, which writes more than half of the hail insurance in Saskatchewan, determines its rates based on its 100-year database for each township.
Schoettler also pointed to SMHI’s business model, which he said is different than how many other hail insurance providers operate. SMHI coverage is tied to property ownership with 80 percent of farmer-clients choosing to renew their coverage in subsequent years through the yearly crop reports they fill out.
Other companies are often represented by agents that represent several insurance products.
Doleman said the AIR model will help Palliser more accurately assess risks, which should lead to better hail insurance rates.
He said farmers in areas that experience large numbers of hail events already pay high insurance rates, but he believes that using the AIR model together with the company’s existing database can result in rates that are more responsive to weather data with insurance rates adjusted depending on if the weather improves or worsens.
The system can also help keep historical biases out of rate assessments, he said. Rates for localized areas hit by heavy hail are bound to go up accordingly if many claims are filed.
However, Doleman said the probabilistic model uses an “elegant algorithm” to determine whether such events are actually only one-in-50-year or one-in-100-year events and if so, could adjust rates to account for relative rarity.
“We need to make sure that we’re not biased by too long a history of data,” he said.
The model will also be able to adjust for changes farmers are making to mitigate hail damage, he added, such as improved farming practices and seed genetics.
Palliser, which is Canada’s largest private hail insurer, can incorporate those beneficial changes into the risk rate process to the benefit of farmer clients, he said.