The insurance industry is concerned with the detection of fraudulent behavior. The number of automobile claims involving some kind of suspicious circumstance is high and has become a subject of major interest for companies. This article demonstrates the performance of binary choice models for fraud detection and implements models for misclassification in the response variable. A database from the Spanish insurance market that contains honest and fraudulent claims is used. The estimation of the probability of omission provides an estimate of the percentage of fraudulent claims that are not detected by the logistic regression model.

Citation: Ayuso, M., Guillén, M. and Nielsen, J. P. (2018). Improving automobile insurance ratemaking using telematics: incorporating mileage and driver behaviour data. Transportation, This is the accepted version of the paper.This version of the publication may differ from the final published version.Permanent repository link: http://openaccess.city.ac.uk/19160/ Link to published version: http://dx. AbstractWe show how data collected from a GPS device can be incorporated in motor insurance ratemaking. The calculation of premium rates based upon driver behaviour represents an opportunity for the insurance sector. Our 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 and which may affect claiming behaviour. We propose implementing a classical frequency model that is updated with telemetrics information. We illustrate the method using real data from usage-based insurance policies. Results show that not only the distance travelled by the driver, but also driver habits, significantly influence the expected number of accidents and, hence, the cost of insurance coverage. This paper provides a methodology including a transition pricing transferring knowledge and experience that the company already had before the telematics data arrived to the new world including telematics information.

No abstract

Most automobile insurance databases contain a large number of policyholders with zero claims. This high frequency of zeros may reflect the fact that some insureds make little use of their vehicle, or that they do not wish to make a claim for small accidents in order to avoid an increase in their premium, but it might also be because of good driving. We analyse information on exposure to risk and driving habits using telematics data from a Pay-as-you-Drive sample of insureds. We include distance travelled per year as part of an offset in a zero-inflated Poisson model to predict the excess of zeros. We show the existence of a learning effect for large values of distance travelled, so that longer driving should result in higher premium, but there should be a discount for drivers that accumulate longer distances over time due to the increased proportion of zero claims. We confirm that speed limit violations and driving in urban areas increase the expected number of accident claims. We discuss how telematics information can be used to design better insurance and to improve traffic safety.

No abstract