On a Clear Day...

It's the sad tale of insurance that incredible pain and misfortune have defined the industry. But out of that suffering have come better ways for carriers to do business and ease the burdens of their policyholders. Events such as 9/11 and hurricanes Andrew and Katrina have a place in insurance history books because they taught carriers new things about exposures and risks, particularly the value of geographic information systems (GIS).

The first significant milestone for GIS came in the wake of Hurricane Andrew in 1992, believes Gail McGiffin, a partner with Accenture responsible for underwriting solutions. "[Insurers] realized they had much greater accountability to manage their exposure accumulation across the organization and protect surplus," she says. "[Catastrophe modeling] became a mandate from the board of directors for most insurance carriers."

The next step, McGiffin explains, came in the aftermath of 9/11, when two things happened. One, it became more important to understand a company's accumulation of risk in a four-wall structure, not just in a ZIP code, county, or state. Also, carriers needed to be able to model loss scenarios beyond just property risk because multiple lines of business were activated in that event.

More recently with Hurricane Katrina, McGiffin points out, there was a further need to determine probable maximum loss from true flood as opposed to wind. The dispute still rages whether wind was the cause of loss in many parts of Louisiana and Mississippi or whether it was flood. "We have flood models, and we have wind models, but we don't necessarily have models that look at both," she says. "The insurance industry is a clear example of these events spurring on the next generation of sophistication analytics and a better leverage of GIS tools."

Besides tracking storms, wind, hail, tornados, and earthquakes, GIS is giving insurers what Stephen Forte calls a deep lens look into risks on the street level in contrast to ZIP code-level geocoding, which has been the standard for a number of years. When you get to the street level, geocoding references data to the street address and location. "In the end, it allows for more accurate and in-depth spatial analysis," says Forte, senior research analyst for Gartner.

Having the street-level data is particularly helpful in the commercial space, comments Forte. In the past, when insurers considered a risk, they typically would send out a loss control team to look at and understand all the hazards and exposures at the site. With GIS, the underwriters get most of that information electronically. "It's nice to know ahead of time all the surrounding exposures for a risk you are thinking of writing and having that tool at the fingertips of the underwriter," he says.

In both personal and commercial lines, Steve Callahan, senior consultant with Robert E. Nolan Co., observes one of the issues that divide sales and underwriting is the raters don't necessarily know the area. "A section of the ZIP code that may be perceived by underwriting as less than desirable may be on top of a hill with more expensive homes," he says. "When you have the tools to get down to a street-level profile, you get a much sharper view of the risk you are assessing. It helps to ensure there is a fair and balanced risk assessment process."

Insurers also have begun to lay their catastrophe models on top of terrorist profiles, so the carriers can map for corporate-level commercial liability insurance and profile the risks based on terrorists, according to Callahan. Another tier in underwriting is looking at risk aggregation and book-of-business analysis so carriers can aggregate their global risks and not overexpose the company in any particular segment of the market.

The Donegal Group developed an automated underwriting system in house called WritePRO for its independent agency force doing personal-lines business, indicates Chuck Ferraro, Donegal's CIO. Two MapInfo products, MapInfo Professional and MapMaker Plus, are incorporated within WritePro to improve quoting, underwriting, and issuance of policies for agents via the Internet.

Agents access the system through the Internet and do real-time rate, quote, and issue; process endorsements; and get a completely underwritten quote back. "Probably 80 percent of the time a human underwriter never sees the submission," says Ferraro. "The GIS comes into play in that it assigns the rate based on the address, rating territory, and property protection class, which are key components in the rating process."

The agent will key in the address, he explains, and based on the data that is coded in the boundaries, the rating territory is defined. "If your rating territory falls into a geographic area that would make it, say, a seven rating, that's going to make your rate different from someone who is in rating territory three," he says.

Dawn Kirchner-King, Donegal's manager of agency automation, reports when the address keyed in by the agent is pinpointed, a geocode–the latitude and longitude–is applied to the territory map the carrier has built into the system, which produces the territory code.

Before GIS became part of the Donegal system, the agent would have to look up the territory location and plug it into the application, adds Kirchner-King. On the back end, the underwriter would have to verify the territory by a paper manual.

Many insurance products are approaching commoditization, and companies agree service will be a key differentiator, suggests Callahan. "The more you know about your clients, the more you can provide differentiation that is relevant to them–whether it is language differences in your service center or higher-profile concierge services for top income levels," he says. "Knowing the product is going to be a commodity means you have to differentiate on service, and that means you need to understand a whole lot more about the people you are serving to provide that differentiation cost-effectively."

As insurers are starting to use GIS in sales and distribution management, carriers are matching agent saturation and competency by market demographics, reports Callahan. This allows carriers to move a higher-profile agent or relocate a higher-producing agent who does better with high-income clients to market segments that are known to have a high level of high-income individuals.

"GIS is taking the retail side of location intelligence and looking at it from an insurance perspective to locate agents to maximize market opportunity without having too close of a proximity to competitors or overloading an area so your agents are competing with each other," says Callahan.

Another area of service, explains Forte, has a carrier's claims department track a hurricane and use its resources to pinpoint its policyholders in the path of the storm, advise them on what they can do to minimize damages, and instruct how to contact the carrier if the storm strikes.

GIS also allows claims departments to be more proactive in other ways, adds Forte. For example, a carrier can study the areas where the highest number of claims is received each year. If the carrier realizes it is getting hit with a lot of claims in a certain region, it can go back to underwriting to push for a better rate. In turn, the underwriters can make a decision not to write in that area, or they can have some hard data for the state insurance commissioner to file for a rate increase and have data to back it up.

There are a limited number of vendors offering GIS solutions in underwriting, Forte points out, and the tools still have to establish their effectiveness. "It's viewed as a good tool to have, but can [vendors] make the business case, and is there good ROI for making this investment?" he asks. "That's where some insurers fall short. They recognize it's important and should be embedded in an underwriting or claims application, but going out and making that cost justification isn't out there yet."

The WritePRO system has been a huge success for Donegal, though, affirms Ferraro. "I've been at this a long time, and [the system] probably has been the most impactful automated system I've ever had the pleasure to be involved in," he says. "It has substantially increased the volume of quotes we make in a given week."

In addition to a notable increase in business, Ferraro believes the system has given Donegal more information. Prior to using WritePRO, quotes were submitted manually, and the Donegal agents would go to a rating disk in their office. "We didn't even know whether they were quoting us," he says. "All we knew for sure is whether they issued a policy. Now, we know they are quoting us, we know what our hit rate is, and we have a lot of statistics about quote activity we didn't have before."

The Georgia Insurance Department began using Maptitude software from Calipers in 2006 to geocode the state and identify latitude and longitude coordinates, according to Donald K. Hill, assistant commissioner and CIO for the Office of the Commissioner of Insurance. When tornados struck southwestern Georgia earlier this year, the commissioner's office was able to put the information it had compiled to good use. "One of the things we needed to do was dispatch our people," says Hill. "If you have damage to a building, our inspectors are trained to determine whether that building is safe to enter." There are other issues from the damage assessment standpoint that had to be considered, explains Hill. "You have to get a fairly quick assessment of the damage from a dollar standpoint," he says. "[Those assessments] kick in how we have to deploy people to get insurance companies in there to start adjusting so people can receive checks and get the damage taken care of."

Georgia insurers are very positive about how the state is assisting them, reports Hill. "The more information they can get, the better," he says. "We were getting hour-by-hour assessments of damage and its extent from people on the ground. We also got photographs [of the damaged area], which we can display on Maptitude."

GIS is just beginning to come into its own, Hill believes. "People are starting to understand how you can take data, geocode it, and apply that data from a layering standpoint into a visual where you can see a lot of things that are happening and how resources can be deployed," he says. "If you have the information or a way of displaying that information in real time, it gives the decision-makers a much better perspective of what they have to do and how to respond."

The cleansing of address data through GIS technology has moved to the front of the underwriting process, according to McGiffin. Mapping also has moved from being a back-office activity to where it now is in the hands of underwriters to help make better decisions and even prospect potential risks. "[GIS] has a role in the transaction side of the business, not just in the management side," says McGiffin. "If we look at the models, we now have additional models for things such as terrorism. We've added to the perils for which models have been created."

As carriers receive an address on an application, the first thing they do is send that address through a cleansing, matching, and geocoding process to verify it is a bona fide address, explains McGiffin. The cleansed address then is used to gather more information or to append some additional information coming from third-party sources. "Now that you have a latitude/longitude as a common denominator, you can go out to multiple sources of location data and request insight on the location and find out more about construction, occupancy, and protection of that location," she says. "You can get more data, verify the data you got from the agent and the customer, and then start a more efficient process to expedite underwriting decisions."

Once carriers have looked within the four walls and know about the building, occupants, and adjacent exposures, the underwriters can assess exposures around the building. Those can be geophysical hazards such as distance to water, distance to earthquake fault line, and where the building is in relation to flight patterns or utility lines. "You need to understand the physical nature and the hazard [those factors] bring to that risk as well as access for evacuation," she says.

Underwriters are getting a visualization of the area, McGiffin reveals, whether it is aerial photography or satellite imagery. "[Visualization] is not a substitute for the mapping data, but it is an additional perspective," she says.

While catastrophe models once were found only in the office of reinsurance brokers, today GIS is moving to general insurers. "There are a number of reasonably priced solutions personal and commercial lines underwriters can use directly as part of their process," says Callahan. "The individual writers have significant exposures as they have learned from catastrophes."

Hurricane Katrina is a perfect example of this need, Callahan points out. A company may have determined pre-hurricane it had 15 percent exposure in the New Orleans area. If it was able to get to the street level, it might have learned 90 percent of that 15 percent was in a flood zone. "That's kind of a different profile than having 15 percent spread out in high-risk and low-risk areas," he says.

Many primary insurers have determined they need to make the investment in GIS, adds McGiffin, although smaller companies will continue to depend on reinsurance brokers for the service. "More of the medium and large insurers are investing in GIS capabilities and want it on an enterprise level," she says. "It's becoming a necessity–particularly on the heels of A.M. Best assessing how companies monitor and manage their catastrophe exposures–so more companies realize this is strategic to their business."