News Currents

For regional carriers, predictive modeling is about more than pricing

New analytics mean more than better pricing; they bring regional insurers closer to their agents and attract new business

Frank J. Coyne, chief executive officer of Insurance Services Office, recently warned that insurers unable to keep up in the technological arms race face a grim prognosis.

The industry leader said a chasm is growing between insurers using state-of-the-art analytics or predictive modeling and those that aren’t – the “haves” and the “have nots.”

“Increasingly, the ‘haves’ will be able to write business at the margins they target,” said Coyne. “The ‘have nots’ will fall victim to adverse selection, as more sophisticated insurers target the risks apt to prove profitable. And the bar will rise continually as competition drives weaker players from the field. Today’s ‘haves’ are at risk of becoming tomorrow’s ‘have nots.'”

National writers such as Progressive and GEICO have grown their auto insurance books for years by virtue of their superior predictive models, but other insurers are now catching up.

John Barbagallo, president of Progressive’s Drive Group, CEO acknowledged as much in a July Insurance Journal interview in which he was asked if his company might be feeling the heat from other personal lines insurers adopting more advanced rating tools. He agreed “the level of sophistication overall is improving. It’s an area where we feel we’ve enjoyed a competitive advantage, and we’ll continue to invest in that, and improve our skills. I think more carriers are getting good at that particular skill set.”

Regional insurers with the resources are adopting the so-called new analytics once reserved for national giants and developing their own segmentation or predictive modeling tools for underwriting and pricing. These complex, often proprietary, computer models employ internal and external data to produce dozens, even hundreds, of pricing and underwriting options for a line of insurance.

The regional insurers like the models not only for their pricing prowess but also because they make it easier for their agents to do business with them. Regional insurers that have adopted predictive modeling also report surges in new business as a result.

Peerless

Keene, N.H.-based Peerless Insurance, an independent agency company that is a member of Liberty Mutual Group, is among the regional carriers that stress using technology to make life easier for agents as much as for improved pricing.

“Ease of doing business is now considered as much a part of our product as the coverage provided by the actual policy,” says Dwight Bowie, president and chief executive officer of Peerless, which writes in the nine Northeast states of Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island and Vermont.

The company’s Personal IQ SmartRisk provides real-time quote and submission capability for automobile and homeowners new business and endorsements. The system crunches internal and external data – credit score being just one variable – to produce more than 40 different price options. In addition to real time quotes, the system offers agents real-time endorsement quotes, policy inquiries, motor vehicle reports and transaction histories.

Peerless is also applying new analytics to small commercial lines. Its Commercial IQ SmartRisk provides up to 100 pricing points for business owners’ policies and simplifies eligibility criteria.

With multiple pricing points, Peerless is now in a position to more accurately price its risks. “We won’t be adversely selected against when we compete,” says Victor Pepin, vice president for personal lines.

Along with the expansion of its pricing has come an “expanded” appetite for new business as well. Agents who used to pigeonhole Peerless as being interested in certain type accounts now ask Peerless to quote on risks they may not have submitted to it in the past.

“Our quote counts have increased dramatically,” Pepin says, adding that “hit ratios” have been good, too.

He maintains the new accounts are coming from a wide range of competitors, including national agency companies.

Pepin believes that its technology puts Peerless in line with the “most advanced regional carriers” and in a position to compete against the Progressive and GEICOs of the world. “We offer the best of both worlds. As a regional carrier we feel we are close to the customer and the local market yet offer the strength of a national carrier,” he maintains.

Hanover

Connection Auto is the multi-variate auto product with customer segmentation capabilities from The Hanover Insurance Group, which includes The Citizens and Hanover Insurance.

Connections Auto has gone like gangbusters. “We had a pretty aggressive plan and our expectations have been exceeded,” James Hyatt, president of personal lines for Hanover/Citizens, told Insurance Journal.

Just one year after its March 2005 release, before it was even available in all of its states, Connections Auto generated more than $85 million of new business. This past summer, the company boasted that business jumped 150 percent in each state where it was introduced.

It has also helped The Hanover add more than 300 new personal lines agents and the company says it has seen a number of “dormant agencies brought to life” since the product was launched.

Connections Auto is now available in 17 states including Arkansas, Connecticut, Florida, Georgia, Illinois, Indiana, Louisiana, Ohio, Oklahoma, Maine, Michigan, New Hampshire, New Jersey, New York, Tennessee, Virginia and Wisconsin.

Hanover says the product allows agents to quote and issue new business for 90 percent of the market. “Companies used to play at the most preferred end of the spectrum but now predictive variables let us reach down into middle tiers,” Hyatt explains.

While Hanover can now reach into middle tiers, it is not interested in building a huge book of substandard business. It is, however, now better positioned to compete where there might be a higher risk factor associated with a good account, such as a teen driver in a family account.

As is true for other regional companies, Hanover values the new analytics not only because they make pricing more accurate but also because they make it easier for agents. An agent can get a quote in about two minutes.

“The older product was unsophisticated like the car I drove right out of high school,” Hyatt says, suggesting the new model is more like a vehicle with electronic stability control and a global position system.

There’s no looking back. Believing that technology is driving competition, Hanover is looking at other ways to apply it. Hyatt says the company is “well down the path on small commercial.”

Selective

Ward’s and Fortune magazines have cited Selective Insurance as among the nation’s best performing companies. Agents consistently rank it as among the easiest and best service companies and it has a case full of awards for leadership in industry technology.

Since it offers just about every technology tool for agents – often before other insurers – it’s not surprising that Selective is also among the regionals adopting predictive modeling. The Branchville, N.J.-based insurer sells commercial and personal lines through approximately 750 independent agents in 20 eastern and midwestern states.

When it comes to the new analytics, the company has taken a broad perspective – viewing these tools as part of an overall knowledge management responsibility. At Selective, it’s about more than rating.

“It’s all about making better decisions how to better use information,” explains Ed Pulkstenis, senior vice president, commercial lines underwriting officer. “It’s more than just pricing. It allows us to broaden our appetite as well as target services more precisely.”

The Selective approach also means using the new analytics to supplement other management information, not replace it, says Sharon Cooper, Selective’s communications officer.

The book of business for the company, which broke the $1 billion revenue mark in 2002, breaks out to be 85 percent small and medium commercial and about 15 percent personal lines. Thus, although it has been pursuing segmentation modeling in both commercial and personal lines, its first application has come in commercial lines. Selective is applying it first in the 20 states where it writes workers’ compensation, a line which it felt to be “most ripe” for the new approach.

“It’s where we were challenged on profitability,” John noted, explaining that a small percentage of accounts were not profitable. Predictive modeling offers a way to do a better job of risk selection at the same time that it promises to improve pricing. It also provides for the targeting of safety and risk management services where they will do the most good.

According to John Marchioni, Selective’s senior vice president for personal lines, while large national carriers may have been ahead of regional carriers in utilizing predictive modeling in personal lines, that’s not necessarily the case in commercial lines.

Selective is utilizing predictive modeling in its New Jersey auto business and plans are underway to expand it to all nine states where it writes this line. Small commercial package policies will be next.

Like other insurers, Selective has found the new tools let it compete against national carriers and in some cases win over new accounts it might never have had a chance with before.

“Agents see a change in our appetite, a willingness to talk about new segments,” Marchioni said.

“We’ve identified some new opportunities more quickly than we might have,” he adds. But there is “always that 10 percent that’s residual business.”

Cooper suggests that “some national carriers have the tech but lost the touch whereas some regionals provide hands-on personal attention. The power lies in effective leveraging of both.”

While small insurers with an expertise or niche have simpler books of business and may be under less pressure to adopt the new tools, for others predictive modeling is a “necessary part of dong business” and expected going forward, says Marchioni.