How GPS information can help improve vehicle insurance premium rate calculation
This research shows how combining traditional motor insurance rating factors with information on driver behaviour made accessible by GPS technology could improve ratemaking and bring benefits to both the insurance industry and its customers.
Standard vehicle insurance ratemaking uses models that predict the expected number and cost of claims based on historical information recorded in an insurance company’s database. Traditionally, the variables included in these predictive models are collected at the time the policy is issued. These variables, such as policy holder’s age, gender, number of years in possession of a driving licence, vehicle power, and location where vehicle is parked at night have known values and either do not change with time or change in a controlled manner. As such, they have their limitations.
However, advances in technology have made it possible to automatically and precisely collect dynamic information about driver and vehicle. Telematics, the technology of sending, receiving and storing information via telecommunication devices, allows information about driver behaviour and vehicle mileage to be collected via GPS-based technology. Information such as the distances driven during a given period of time, and also driver habits and behaviour that may undergo changes during this time, can be captured.
Employing real data for their study, researchers from Riskcenter, University of Barcelona and Cass Business School have tested the combination of traditional motor insurance rating factors with new information obtained from telemetric data collection. Their approach is based on count data regression models for frequency, where exposure is driven by the distance travelled and additional parameters that capture characteristics of automobile usage.
The results show that variables related to the annual distance driven and to a driver’s behaviour lead to better estimations of the expected number of accidents than those reached when using the traditional variables. However, the model that performs the best is the one that includes both traditional variables and the new telemetric variables, with the annual distance included as either a regressor or offset (risk exposure) in the model.
This research demonstrates that usage-based information is informative for premium ratemaking, and that the combination of classical actuarial insights with telematics information is superior to working with either system in isolation.
The calculation of premium rates based upon driver behaviour represents an opportunity for the insurance sector. Telematics technology offers potentially improved actuarial accuracy and may benefit those policyholders who drive less. It has the potential to modernise the ratemaking system and bring about fundamental changes in automobile insurance in the near future. It seems clear that future ratemaking models should and will incorporate these technological advances.
A post-peer review edition of the paper Improving automobile insurance ratemaking using telematics: incorporating mileage and driver behaviour data is available for download at the link below. The final version was published in Transportation in June 2019.