Keeping a vital practice alive: Indiana Lumbermens Mutual Co. recognized that without a process in place for improving risk selection, the company would continue to struggle in today's economic climate. Given its situation, ILM knew it had to explore risk assessment technology solutions

Risk & Insurance, May, 2004 by Dave Walters

An inconsistent investment environment and changing loss patterns have forced insurance carriers to focus on the foundation of their businesses: underwriting. According to A.M. Best, only a small percentage of insurers have managed to keep combined ratios below 100 percent, leaving the vast majority of the industry operationally unprofitable.

Indiana Lumbermens Mutual Co., a provider of property and casualty insurance to the building materials and forest products industry, was no exception. Facing increasing severity of losses in its workers' compensation division, ILM needed a method for improving risk selection ability. Terminating all workers' compensation business would have a large impact on ILM's overall business, as they tend to write all lines of coverage for many multi-location accounts.

By implementing a process for yielding better risk selection, insurance carriers can not only improve underwriting profits, they can retain an edge over their competitors and remain stable in varying economic climates. ILM recognized that without a process in place for improving risk selection, the company would continue to struggle in today's economic climate.

Given its situation, ILM knew it had to explore risk assessment technology solutions. Within the workers' comp division, ILM recently completed a proof-of-concept process with Valen Technologies' Risk Manager platform for underwriting decisions. The detailed assessment proved that the risk assessment platform could improve ILM's loss ratio by as much as 9 percentage points, based upon individual class/ location studies. Given that ILM's workers' comp line has been trending negatively, a 9-point reduction would have a striking impact on their profitability.

A CUSTOMIZED PLATFORM

A risk assessment technology platform can be equated to arming all of a carrier's underwriters with complete and total recall of the book of business and an analytical capability to reason among every risk factor. As ILM is more than 100 years old, the carrier needed a solution that would enable its underwriters to leverage its rich, existing historical policy and claims data. At the same time, ILM needed to prepare for the future and be able to conduct profitable business in the rapidly changing economic climate.

ILM chose to test the effectiveness of Valen Technologies' Risk Manager platform, which gives insurance carriers a tool for making highly accurate risk decisions on a policy-by-policy basis. ILM received a customized risk model based on all of its historical data. The model can be used to produce more accurate risk recommendations for new and renewal policies than underwriters' personal recall could possibly produce. Risk assessments made by the Risk Manager model are more accurate because the software considers hundreds, even thousands, of risk factors to uncover all possible predictors. Underwriters can use the model to make decisions on a unit at risk or policy-by-policy basis, and no longer have to exclusively rely on underwriting guidelines based on past experience or subjective estimations to make risk assessments. Given these new technical capabilities, ILM set out to test the proposed solution for its problem.

ILM employed the proof-of-concept process to ensure that a risk assessment platform would be effective in reducing their loss ratios. The process entailed looking at ILM's underwriting decisions in 2001 as a basis for determining whether a reduction in loss ratios would have occurred if the carrier had used the model for those decisions. To begin the process, the dataset was prepared using ILM's policy, claims and loss control data, as well as supplemented data from outside sources. Once the dataset was prepared, a statistically valid portion of the dataset was utilized to build a risk model that would eventually be used by underwriters. The last step in the process was to objectively validate the risk model with unseen policies that were not used in the model development process.

Although ILM's dataset is rich, supplementing it with external data sources would make platform predictions more accurate by allowing more possible risk factors to be taken into account. Leveraging external data sources would also make the platform more relevant to the greater economic climate outside of ILM. After conducting a detailed analysis of ILM's existing policy, claims, loss control and underwriting data, external data sources were identified that would complement internal ILM data to achieve a more accurate predictive model representative of all available workers' comp predictors. The dataset was supplemented with external data from the Bureau of Labor Statistics and the Occupational Safety and Health Administration, among other sources. Finally, all risk factors to be used in building the model were defined.

To ensure that Risk Manager directly addressed ILM's root problems, the carrier established its financial metrics for acceptable and unacceptable risks. ILM determined that it would prefer not to write labeled policies with a loss ratio over a defined percentage.