North Carolina is consistently in the top 5 for the LOWEST car insurance rates in the nation. But how do insurers arrive at these rates?
Insurance companies are mining a vast amount of data about you — from your shopping habits to credit history — to help calculate premiums.
Are you a bigger risk if you wait until the last minute to buy car insurance, than a customer who purchases it a week or more ahead? Are 16-year-old drivers in Fort Lauderdale, FL more or less likely to wreck their cars than 16-year-old drivers in Seattle, WA.? Does a history of making many auto insurance claims also mean you’re likely to make homeowners insurance claims?
These are among numerous questions insurers are contemplating as they look for new clues to assess risk. The process is part of a changing science called predictive modeling, which has revolutionized how the industry prices auto and homeowners insurance policies.
Steven Armstrong, a fellow of the Casualty Actuarial Society and the global chief actuarial officer for consumer insurance at Chartis Insurance says “Its intent is to predict the future by making mathematical sense of the past,”.
Predictive modeling uses statistical and analytical methods driven by technology to sort through millions of data points. It searches for patterns and clues that point to how likely customers are to make claims.
The mathematical theories that render predictive modeling possible have been around for many years, says Eric Huls, a vice president of quantitative research and analytics at Allstate. “But there wasn’t a machine powerful enough to apply them to large data sets. What’s really advanced is computer power.”
Predictive modeling has really increased in use in the last decade due to great competition between insurers, increasingly available data, and tech advancements says Claudine Modlin, a senior consultant at global research and consulting firm Towers Watson.
“Today, data is literally everywhere,” says Chartis’ Armstrong.
Information not available to insurers directly from you can be purchased from vendors. Armstrong says he’s heard of predictive modelers searching data about what people purchase at retail stores to look for any connections between shopping habits and risk. He’s not aware of any insurance firms actually using that info to set rates.
Details, details, details….
Predictive modeling provides insurance companies the ability to review an enormous number of variables in more combinations than they could 25 years ago. Back then, an insurer would likely price an auto insurance premium based on 15 variables, Armstrong says. “Today 40 or 50 variables might be considered.”
Insurance companies are also using data with more detail, including customer birth days versus broad age ranges, Towers Watson’s Modlin says. Some home insurers are considering how features of the homes and neighborhoods such as number of bathrooms, square footage, and whether the plumbing or electrical has been kept current might predict claims for losses.
In the past, insurance companies have considered that type of data to ascertain how much home owners coverage might be needed. Now they’re considering how this info might correlate with risk. For example, does the median age of people in the community correlate with theft risk? What about the community’s employment rate?
Predictive modelling also enables insurers to isolate the correlation of each variable with risk and then analyze how that correlation could be different in specific circumstances, Allstate’s Huls says.
As an example, an insurance company can consider how the combination of where you reside and your age correlates with risk. Insurers can set premiums that more accurately reflect the chance you will file a claim with the ability to isolate variables, as well as examine many combinations of them.
All of this collection and review of detailed data about you may seem invasive, but improving pricing accuracy benefits the customer, as well as the industry, Modlin says.
“Nobody wants to pay more than they have to, and at the very least they want it to be commensurate with their risk,” she says. “You don’t want to be a good driver who’s paying more to subsidize a bad driver.”
Prior to predictive modeling, rating structures were similar between insurers, Modlin notes. Now, the structures are a great deal more sophisticated and complex, and it’s not as simple for one insurer to understand another insurer’s rating methodology.
With varying rating structures, it’s always a good practice to get multiple quotes for your coverage. You can also speak with your agent to review ways to reduce your premiums. Saving on insurance is by comparative shopping is one of the easiest ways to save money without loosing value…It’s usually well worth the extra time invested.