Predictive modeling gives our underwriters the confidence to compete on the most desirable business and the flexibility to adapt
rapidly to changes in market conditions. We paved the way for our return to underwriting profitability by collecting more data,
performing better analysis and building better models for pricing policies, evaluating reserves and forecasting financial results.
For workers’ compensation, we began in the second half of 2012 to use our second-generation predictive modeling tool.
For auto, general liability and property coverages in commercial package accounts, we have used pricing precision tools since
late 2011.
For small business policies written through our CinciPak™ product, similar tools became available in several states
in 2012 and will deploy for many more states in 2013.
For personal auto and homeowner policies, we continue to enhance model attributes, expanding the number of pricing
points and precision. Predictive modeling for dwelling fire policies will begin in 2013.
