Even as the U.S. economy improves, insurers continue to face a host of challenges, including long-term low-level interest rates, which dampen earnings and the commodification of insurance products, increasingly sold online at a discount — all of which erode profit.
In the face of these challenges, insurers as diverse as AIG, Swiss Re, Progressive, Falls Lake and others are looking to leverage the power of analytics and big-data sources to be more selective about the risks they assume and price them more accurately, even as they grapple with new and unfamiliar data and technology and the cost of acquiring that technology and institutional knowledge. “Insurance is notorious for looking in the rear view mirror in order to drive forward; that’s the essence of our business,” says Steve Hartman, president and CEO of Falls Lake Insurance Companies: “We make an estimate each time we price a piece of business; we won’t know what that cost actually is for a few years.”
Stonewood Insurance, a subsidiary of Falls Lake, has a 12-year operating history, a stable policy administration system from Tropics Software, and an unusual strategy for implementing analytics, Hartman says.
“We ended up with a hybrid approach. Our data is pretty strong, but it’s not statistically strong. As you think about credibility, more data points are always better than less,” Hartman says. “So, [we implemented] the Valen Manage product, we essentially score our history against the couple of million data points that they have. That allows us then to benchmark with much more credibility to the business that we target.”
One of the reasons Stonewood doesn’t use the Valen data points exclusively, Hartman says, is that it is multi-state and multi-class.
“Our book of business within Stonewood is very focused,” Hartman says. “We are active in four states and we write business in six business segments, which represent about 85% of what we do. What we didn’t want to do was supplant the underwriters’ judgment by introducing black and white rules around usage of a predictive model.”
Instead, Valen offers a frame of reference around adequate pricing and empowers the underwriters to exercise their judgment. “In essence, it’s redefining unity. The predictive models help to define unity based on the account in question, held up against a backdrop of a couple of million data points in the Valen model. We’re trying to frame that as the reference point for the underwriters as they think about the combination of exposure, policy form, and pricing.”
In addition to improving underwriting results, Hartman – who did his graduate studies in cognitive psychology — says this implementation is successful in part because it respects the underwriters’ knowledge and is sensitive to their concerns of being replaced.
“Underwriting is, in essence, taking less-than-perfect information from a variety of sources and trying to put a puzzle together,” Hartman says. “As each person approaches the same information, the pieces of the puzzle look different. Usually that’s determined based on guidelines within the company, but it’s also determined by individual bias: past experience and that kind of thing. We’ve tried to refle t the relative of importance of the individual’s perspective, which can’t always be quantified, then marry that to a model that is driven by quantified ariables.”
Stonewood has been using these tools and methods for about a year, and is able to roll up scoring across the portfolio, and look at agency performance, as well as class, state and line of business.
“We’re measuring the difference between model output and achieved pricing. It doesn’t mean that one is right and the other is wrong; it’s simply the case that we want to make sure we understand where the differences are,” Hartman says. “Then we do some additional analysis through our actuarial group to make sure that we understand those differences and can quantify any impact from them.”
Hartman says the commitment to underwriting analytics is both an offensive and a defensive investment. “On a defensive basis, as we go through the normal market-cycle rates’ ebb and flow, we want to make sure we are best positioned to defend the renewal base we want to maintain. On an offensive basis, we need to match exposure to contract to price. I want them to understand exposure firs , then apply the contract form that we provide to make sure there are no disconnects. The third piece of that equation is determining an appropriate price. That requires a really good staff with a good toolbox and the ability, at a company level, to roll it up and analyze on a directional basis where we’re heading. And, separately, to analyze the expected performance on that business.”
Hartman says insurers must routinely invest in and improve their analytical toolsets to ensure that they are picking the best business, minimizing credit risk and making certain their claims operations are efficient. “These are the markers that are embedded in your business and can indicate whether you may want to turn left, right, or step on the gas. If you’re not paying attention to those signs, it’s going to be easy to find yourself at the bottom, wondering what happened.”
Insurers are tackling underwriting analytics in myriad ways. Check out their stories:
Register or login for access to this item and much more
All Digital Insurance content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access