Creating an Auto Insurance Pricing Model for a Deregulated Future

A P&C insurance company based in the Asia-Pacific region faced the prospect of intense pricing competition and lower profits due to the imminent deregulation of the local auto-insurance market. BCG used innovative and robust methods in order to help the company understand the impact of multiple variables on risk and loss ratios and price its products as competitively and profitably as possible.

In contrast to a system that required insurers to use unified premium guidelines and price their policies at similar rates, the new regulatory regime in the host country—typical of the deregulation now taking place throughout the region—will give companies significant freedom to differentiate among customers and price their products accordingly. Like most insurers in the region, this P&C company had previously assessed risk with a simple methodology that did not take the interdependency of variables into account, thus limiting the degree of granularity in customer risk profiles.

Using existing policy and vehicle data, BCG input 20-plus variables—including two that the company had never before used: postal code and purchasing power (the latter obtainable from commercially available data sets)—into a generalized linear model (GLM). The GLM (which is simply a method for analyzing the interdependency and impact of multiple variables) was then run to predict risk and cost from the frequency and seriousness of claims associated with individual customers.

The impact analysis showed where there were opportunities to lower premiums for what the model identified as lower-risk customers and to increase them for higher-risk ones. BCG set up a data team to create visualizations of the impact analysis generated by the GLM, to help the actuaries communicate the findings to management effectively.

In addition to developing and implementing the GLM to produce the impact analysis, BCG used the results of the analysis to design new product packages—including policies customizable for individual customers—for the company to test in the market. We used the GLM and competitor data to see whether the company’s pricing is in line with the market. We also worked closely with the data team to embed GLM capabilities in the organization for making future pricing decisions.

Impact of the New Pricing Model

With its new pricing model, the company has acquired a true understanding of the risk cost associated with individual customers, agents, and business partners.

protected by reCaptcha

Subscribe to our Insurance eAlert.