Did You Know? Auto and Renters Risks Correlated

February 4, 2020 by Susanne Sclafane

To Alex Timm, CEO of Root Insurance Company, the auto insurance carrier’s recent move into the renters insurance market seemed like a natural extension for its customer base given that 70 percent of its policyholders are renters.

“We actually had a lot of people asking for it,” he said.

In addition, the auto insurance company already had a key piece of data to price the business.

“If you look at one of the leading variables in most of the competitor pricing plans, it’s actually the prior losses on your auto that they’re looking into to basically figure out how much they should charge you for renters insurance,” Timm said, explaining how Root can offer renters insurance at such low rates—as low as $6 per month, according to Root’s November announcement about the new offering. “We know that the auto risk is very correlated with the renters risk, which allows us to then have a bit of a deeper understanding,” Timm said.

The reverse may be true also, according to a paper published by researchers from the University of Warsaw (Kinga Kita-Wojciechowska and Łukasz Kidziński) in April 2019. The paper, “Google Street View image of a house predicts car accident risk of its resident,” (arxiv.org/abs/1904.05270), also the subject of an article in MIT Technology Review, found that features visible in a Google photo of a house—type (detached, terrace, etc.), age, condition, estimated wealth of its residents, density of surrounding buildings—can be predictive of car accident risk, independent from classically used variables such as age or ZIP code, and can also increase the predictive power of traditional auto insurer models.

For Root, there are considerations beyond the availability of rating variables as the carrier enters the market.

“It is a low average premium product. So, really, the underwriting risk you’re taking on renters is minimal compared to other products,” said Timm. A more pressing question, then, was whether Root could acquire customers at a good customer acquisition cost. “For us, because we already have these mobile customers”—the auto insurance app is right on the mobile phone— “we can make it super simple for our consumer to purchase,” he said.

Timm said Root also increases its retention in that book of business when it cross-sells. Renewal loss ratios are “much, much better” than its new business loss ratios and “so it also boosts the overall profitability,” he added.

In the auto insurance world, Root’s mission centers on the concept of fair pricing using driving experience. By offering a mobile-based telematics product for good drivers, Root boasts the ability to slash some auto insurance bills in half.

Root may not stop at auto or renters. What about full homeowners or life insurance, which are now being vetted by established players like Progressive? Or commercial auto, a seemingly natural play for a telematics-based insurer?

“We are always evaluating those opportunities. Right now, there’s nothing near term that we’re planning in that regard. But we do believe that being a one-stop shop for consumers and always offering them the best product, whether that’s our product or whether we’re brokering it for somebody else, we believe that that’s really where the future is going to be,” Timm responded.

Coming to a State Near You

Root renters is currently available to existing Root auto customers in Missouri, Ohio and Utah, and will be expanding to other states. Insurance will also be available to renters who are not Root auto customers in the coming months, the carrier said.

On the auto side, Root was in 29 states as of August 2019, with plans to be nationwide by 2020.

Nationwide? Aren’t there any states that are off limits because of congested conditions that impact driving behavior or other factors?

“We’ve really found, regardless of the area of the country, even if you’re in New York City or Chicago, for instance, that there’s still a large percentage of people who are driving better than average for that cohort, which means they deserve a discount,” Timm said. “That being said, there are still regulatory challenges throughout the states. Michigan, for instance, has a regulatory issue with the no-fault laws there. So, that’s something we’re not prioritizing.”

He said Root is now working with regulators in New York. “We believe this is something that, in the state of New York, would be great for their consumers and there’s a lot of different segments too to the New York market that we believe we offer a really unique solution for.”

California, a state where Root insurance is already available, also poses some special problems.

“We aren’t aggressively pushing the product or marketing [in California]. They don’t allow telematics, which is very unfortunate because we believe this is something that obviously really helps consumers, particularly consumers in traditionally discriminated-against areas and classes. We believe this is something that absolutely is good for the consumer. We plan to work with California to make sure that the consumers can be a part of this because it is a very big benefit to the consumer. It’s much fairer than what typically is used.”

“It’s kind of a tragedy that it’s not allowed,” said Timm, reminding Carrier Management of a younger version of another actuary, J. Robert Hunter, in his dedication to consumers.

Hunter recently wrote a letter to his colleagues in the Casualty Actuarial Society, criticizing their white paper, “Insurance Rating Variables: What Are They and Why They Matter. “Your paper says nothing about the plight of good driving, low- and moderate-income Americans being unable to afford state-required auto insurance due to the negative impact of socio-economic, non-driving-related rating factors on these people,” Hunter wrote.

Likewise, Timm has said that existing pricing models for auto insurance, based on variables tied to economic status, essentially make the product a “tax on the poor.”

Do poor people use the Root app and qualify for Root insurance? Has Root solved the problem?

“To a large extent, yes we have…The beauty of telematics is that it allows every consumer, regardless of their credit score, their age, their gender, their marital status or ZIP code really, a chance to prove, ‘Hey, I’m a good driver. I deserve a better rate, even if my credit score isn’t the best, or even if I live in a suboptimal area.’ And that’s really what our pricing model does,” he said.

What about fraud? Does anyone cheat on the driving test by asking better drivers to take it for them?

While there are ways to trick the system, Root has found “that most people are very much married to their phone [and] won’t give it away for a long period of time,” Timm said, noting that Root is also able to flag unusual behavior. “If you give it to somebody and they drive just for one drive, but every other day of the week the phone’s not going to that location, well that’s odd. Or, if you give your phone to somebody and it looks like where they’re driving and keeping the phone isn’t the same as your address,” he cited as another example.

“It’s not going to be completely perfect, but what we’ve found is that really the proof is in the pudding,” Timm said. “In terms of predicting frequency of auto accidents, we are seeing that our internal telematics platform is hands-down the most predictive of any of the variables that we use right now.”