Only 30 years ago, pricing for personal-lines insurance products was simplistic and comparatively primitive. Desk underwriting was almost universally used for the application of rules and risk management, due to the fact that foundational analytical techniques were insufficiently accurate.
Consequently, a staff of experienced professional underwriters was essential to a firm’s profitability. Pricing was more influenced by heuristics based in common knowledge, actuarially developed univariate relativities, and on urban legend. These risk models produced a handful of price points, and it was up to the underwriting staff to segment the universe of risk, selecting only those risks that would hopefully produce acceptable results.
The inefficiencies were multidimensional. The risk models would support only narrow fields of view of the market, thereby limiting a single company’s ability to grow. Desk underwriting represented high frictional costs and the inconsistencies associated with many human beings interpreting paper manuals differently introduced uncertainty in the data environment. “Hard” and “soft” market conditions, agency relationships, and company objectives all contributed to a certain fluidity of underwriting discipline across the industry.
Today, numerical techniques and statistical frameworks are more commonly used. Predictive modeling is now common practice, even beyond pricing analytics. We are seeing linear, nonlinear and mixed-model forms being utilized along with innovations in logical structures and experimentation with new variables and data sources.
Well-designed models will compensate for interactions and correlations among rating and classification variables, and the number of risk classes and price points allow insurers to automate business acquisition and maintenance processes almost completely. Today’s analytical methods enable an insurer to market to almost the entire spectrum of risk with greater confidence in the financial outcome. The drill is no longer “accept or reject”; it is now “classify, rate and issue.”
The remarkable advancement in technology we’ve experienced over the last few decades has been the favorable wind in the sails of analytics. Thirty years ago, technical infrastructure was almost completely dedicated to policy administration, billing and print functions. Analysts and actuaries now routinely run computationally intensive applications and procedures that were sheer fantasy a few years ago. Sufficient processing power is easily within the reach of even the most modest IT budget, and the cost of magnetic storage seems to drop every day. That leads me to the following conclusion: Applied analytical methods will become an increasingly important component of successful organizations and a point of differentiation among companies.
Companies like Narragansett Bay Insurance Co. realize that leading-edge technical infrastructure, data and enterprise-level analytical cultures are strategic assets. Legacy best practices are giving way to advanced techniques like sequenced modeling that combine parametric and nonparametric methods, producing risk models that are significantly more predictive.
Consumer behavioral components are more commonly added to the overall equation, leading to more satisfactory ownership experiences for consumers and improved economics for insurers. In response to the enormous differences in threat environments, new geo-spatial territory models are replacing traditional territory boundaries like zip codes and counties; these models offer great promise for both auto and home products.
Homeowners’ and other property products deserve the same level of investment and innovation that has been applied to private passenger auto. Therefore, I predict that the most significant changes in personal-lines insurance in the next five to 10 years will be in homeowners’ and property lines of insurance.
Just as space exploration brought advancements in other disciplines, the insurance industry can learn a great deal by keeping current with other fields. Advancements in biostatistics, analytical chemistry and geography, for example, often have applications in insurance. Analytics, data, and systems are increasingly important in the insurance industry. Winning companies will invest wisely in these areas, as the economic efficiencies they help bring about produce benefits for the consumer and the company alike. •
Jose Trasancos is senior vice president of research and product development for Pawtucket-based Narragansett Bay Insurance Co.
Seifert Systems Invests in Energy Efficiency to Strengthen Operations
For manufacturers, energy is more than just another operating expense. It plays a critical role…
Learn More
