Designer Selections: Predictive Analytics in Insurance

Underwriters face two kinds of pressure, explains Donald Light, a senior analyst with Celent. One is the need to secure more business. The second is to make sure any business they write is profitable.

“Those often are contradictory pressures,” he says. “The easy way to get more business is to offer really cheap prices, but the best way to make sure your business is profitable is to maintain your pricing stance and not make concessions.”

Those pressure points are inflated by what Light calls “a prolonged recession or an anemic recovery, whichever phrase you prefer.”

One problem facing the industry is the amount of insurable business has decreased as the economy has suffered.

“There are fewer businesses, fewer employees, fewer buildings, and fewer products being produced and sold,” says Light. “The pressure to try and maintain the top line–get business on your books–is pretty severe.”

Claudine Modlin, senior consultant at EMB, maintains underwriting and pricing have to go hand in hand. As pricing gets more sophisticated, the underwriting strategy has to adapt accordingly, she adds.

“Underwriting picks up where pricing lets off,” says Modlin. “Fifteen years ago there was a lot of subsidization in the rates and underwriting was designed to allocate risks in light of that subsidization. Now that we see rating plans become much more accurate, underwriting has to adapt.”

When bad times come, all parts of the insurance food chain–insureds, agents, carriers–are affected, explains Paul Zamora, vice president of underwriting for ICW Group.

“You have to look at how you can do things more efficiently,” he says.

For carriers, that means automating as many processes as possible, particularly some of the minute tasks performed by carriers and agents–data entry in particular. “How do we automate [those tasks] so we can do them more efficiently and process more business without additional costs?” asks Zamora. “Insureds are saying they can't afford to pay more in premiums–they have to find ways to cut expenses. Insurance premiums are one of the areas where they are looking at to save money. [Carriers] have to make sure we have competitive prices and in order to do that we have to automate as much as possible.”

Commercial Lines

The greatest challenge to automation has come in the commercial lines, points out Modlin. Commercial lines traditionally have been high-touch in terms of underwriting activity with more wiggle room for the underwriter.

“Now what we are seeing in terms of underwriting strategy is either more discipline on the manual filed-rate side or more discipline on the scheduled-rating side,” she says.

Actuaries might be using statistical models to fit the rates and there are underwriting models that use a variety of explanatory variables and are essentially unconstrained to determine what the best rate might be, explains Modlin.

“The underwriting strategy should be designed to pick up the difference between the two,” she says. “If the filed rate is $1,000 but the underwriting model suggests it should be $1,200, that tells you right there that is a piece of business that is underpriced. What can underwriters do to adjust for that? Can they apply a debit in terms of scheduled rating debits? Can they choose to put it in a different company–more of a nonstandard company? Those are the ways underwriting strategy has changed–the discipline around scheduled rating plans and the discipline around modifying the filed rates.”

Today's Market

Two strategic choices are affecting today's market, according to Karen Pauli, research director for TowerGroup. First, there are carriers that made the decision to keep competing for business no matter what. Secondly, there are carriers that decided not to play that game.

The second group “is not going to get into a competitive bid for business,” says Pauli. “They are going to hold to their underwriting standards and pricing guidelines and if they lose ground, well, they lose ground.”

Pauli believes smart carriers are the ones that discovered who their best producers were, what their best class of business was, and had the technology to focus on pricing strategies that were specific, finite, and focused–both geographically and with the producer in mind.

“You have to be disciplined to pull off strategies,” she says. “The analytics is the key piece right now to assess your best risks and where you have some capabilities to up the price of some coverage without blowing the customer out of the water.”

Zamora feels ICW Group has taken major steps over the last two or three years to prepare for the inevitable return to a hard market

“We recognize that in order to be successful as the market starts to harden you have to have a competitive price,” he says. “We've already laid the groundwork from a technology standpoint. We are operating about as efficiently as possible.”

Zamora also believes the hard market is around the corner.

“I originally thought it would be the fourth quarter of this year, but I'm now predicting it will be the second quarter of next year,” he says. “We're seeing carriers increase prices in California. We're seeing carriers produce results that are not profitable. There is less reserve for them to dip into to make their financials look better. Carriers that are going to thrive will have to operate in a highly efficient manner. We're there.”

Light doesn't agree with Zamora's assessment. He believes the market remains soft, with a few exceptions, such as offshore oil rigs and the energy sector. “Much of that is driven by the lack of a robust economic recovery,” he says.

Underwriting Trends

From a technology standpoint, Light sees three trends prevalent in the underwriting field:

o The development of underwriting workstations

o Automation of tasks within the underwriting process

o The use of analytics

Celent analysts are finding more interest among insurers in moving to the next generation of underwriting workstations.

“In one sense, any underwriter that does any part of the underwriting job with a computer has an underwriting workstation,” he says. “Insurance companies are seeing the value of having an integrated set of screens and capabilities, both for inquiry and for seeking and processing data all in one place with coherent navigation and the ability to share information within the company, with the producers, and sometimes with the insured parties themselves.”

Some vendors have developed underwriting workstation products, but Light points out many policy administration solutions are giving new business functionality much more attention, which he feels translates into some form of underwriting workstations.

Automating Tasks

One area that has helped insurers improve decision-making is with the emergence of business intelligence solutions, according to Light.

“Business intelligence can be thought of as occupying a kind of middle ground between conventional reporting and the analytic process,” he says. “Business intelligence will borrow from both worlds and try to show a lot of data that can be manipulated to some degree and drilled into to some degree by the line underwriters and different levels of management. The fruit of applying analytics to the underwriting process to some degree can be garnered by the intensive use of business intelligence systems. Small or midtier carriers might be steered to a business intelligence system where prices have come down and things are more accessible.”

Modlin sees more automation in underwriting today, particularly with business rules engines. “There always will be a bigger role for underwriter intervention on the commercial lines side, particularly for larger risks. But for some segments of the portfolio–even some like commercial auto that almost smell and feel like a personal lines risk–I could see there being a lot more automation and more business rules engines to make the entire process more low-touch from the underwriting perspective and saving some of their time and energy for larger accounts.”

Modlin doesn't believe the role of the underwriter will diminish in the area of the larger account risks that are more individually priced.

“[Underwriters] have to be involved in the monitoring of this business,” she says. “Are the business rules working as designed in terms of what business we are converting, what frequencies and loss ratios are being achieved? There has to be a stringent monitoring system in place once these rules roll out.”

Business rules engines “have a really good place on this earth for things like knockout rules and triaging business so the underwriter can really focus on the best things,” says Pauli.

She believes every carrier needs business rules engines, but she points out they are backward looking.

“You adjust them based on having reviewed your last quarter's losses,” she says. “You bring your top senior underwriters together and craft the rules and put them in. They are really important, but they are not as good as the analytics models.”

Predictive Analytics

Light feels the use of analytics in underwriting is at an early stage.

“Underwriting decisions are made by some combination of underwriting policies and underwriter judgment,” he says. “Analytics is viewed as a third force in how different underwriting decisions are made. An analytic toolset can be applied to a large amount of data and give insights into the characteristics of a risk being submitted for coverage.”

Even tier-one carriers still have to have the analytic toolsets, the people who use the tools, and the willingness of senior management to incorporate the insights that are obtained.

“The difficulty for smaller companies, unless they tap into outside data sets, is they may not have enough data or information to apply the tools,” says Light. “Plus, [the tools] are relatively expensive and you have to have the staff that can use them.”

Underwriters need to be intimately involved in the analytical process, points out Modlin.

“You can have statisticians or actuaries fitting models involving all these variables, but the underwriter sniff-test ability is invaluable,” she says. “They need to be part of the analytic process in designing these business engine rules and to what risk the rules will apply.”

Pauli believes larger carriers have a distinct advantage in the use of analytics. Many of them have successfully used analytics in the personal lines and are bringing them into the commercial lines area now.

“It's the larger carriers that have been able to do that,” she says. “They have capabilities. They have data. They have IT people, and they can focus on that. Probably one-third of the carriers can do that and they are the bigger folks.”

It's not the cost of analytic tools that is the holdup, explains Pauli.

“What I think is the general perception and to some degree reality is you have to have your data in good order,” she says. “You have to have aggregated data and a data warehouse.”

Pauli believes analytics can help because there are some processes that can be automated and there are some issues where the underwriter needs decision support.

“It's hard to slay the dragon of tradition that commercial lines underwriting is an art form,” she says.

Straight on Through

Most carriers have developed straight-through processing capabilities where an agent can get online, answer some questions, and get a quote within minutes, explains Zamora. This means carriers have to become more efficient in processing business to stand out from the competition.

“You have to make it easy (for the agent and the insured) and straight-through processing has changed that substantially,” says Zamora. “Technology has allowed us to be much more efficient in managing our processes and has allowed us to turn the quote process around quickly.”

At one time, ICW Group had a much larger data-entry staff because the carrier would get most of its policy applications either through e-mail or fax and had to rekey all the data, explains Zamora.

The problem was compounded by internal systems that were unable to communicate the different aspects of the process with each other.

“On our side, all those systems communicate,” says Zamora. “We don't rekey data so we can operate more efficiently with a smaller staff. Our procedures have changed substantially in terms of how we manage our day-to-day workflow.”

Zamora credits the underwriting solution from FirstBest–which ICW Group has branded Snap–with giving the carrier a competitive advantage.

“We are the only carrier in California that has this upload functionality from a workers' comp standpoint,” he says. “We can upload the data through [the agents'] agency management systems, so we don't require agents to go online to our system to rekey the data. We have streamlined their workflow and we can offer a quick turnaround from a quote standpoint.”

Price Optimization

Modlin believes the industry is seeing more pricing sophistication on the personal
lines side.

“Whether that is the use of better methods–those that are more statistically based–or whether it is tapping into additional explanatory variables, I think those are enhancements,” she says.

Modlin also sees better treatment of what she calls the high dimension variables, such as getting the right price by geography.

“There might be hundreds of ZIP codes in a particular state,” she says. “How do you combine them in the right manner to get better territorial definitions?”

Another area that excites Modlin is the use of price optimization techniques. Actuaries involved in pricing business have developed what Modlin calls a cost-based rate.

“There's always been confidence around that rate, but it's the business' job to tweak that rate in a certain direction based on competitive pressures and not wanting to overly dislocate the existing portfolio,” she says.

Goal

What some carriers have done is pass on work to the agents, a route ICW Group has purposely avoided.

“We found a way to use network functionality so both [the agent and the carrier] can be more efficient,” says Zamora. “If carriers are going to be successful in the future, they have to offer a competitive price and to do that they have to be more efficient through technology.”

ICW Group is looking to accomplish that by putting its best underwriters together, picking the biggest class of business, and finding the best approach to underwriting. Zamora feels rules engines and decision-making technology allow ICW Group to be more consistent in its approach to underwriting, but that has brought a change in the responsibilities of underwriters.

“The old-fashioned underwriter was more like a goalie: How do I keep the carrier from making a mistake?” explains Zamora. “Now our underwriters are more focused on sales–building relationships with our agent and broker partners. They are spending more time analyzing risk and less time with the data entry stuff.”