Geographic ratemaking with spatial embeddings

TL;DR

This paper introduces a novel geographic ratemaking method that uses spatial embeddings and features to improve prediction accuracy and enable rate setting in areas lacking historical loss data.

Contribution

The paper proposes a data-driven spatial embedding approach for geographic ratemaking, reducing bias and variance compared to traditional spatial interpolation methods.

Findings

Produces more accurate risk predictions than bivariate splines.

Enables rate estimation in regions without historical loss data.

Offers a flexible, non-parametric alternative for spatial risk modeling.

Abstract

Spatial data is a rich source of information for actuarial applications: knowledge of a risk's location could improve an insurance company's ratemaking, reserving or risk management processes. Insurance companies with high exposures in a territory typically have a competitive advantage since they may use historical losses in a region to model spatial risk non-parametrically. Relying on geographic losses is problematic for areas where past loss data is unavailable. This paper presents a method based on data (instead of smoothing historical insurance claim losses) to construct a geographic ratemaking model. In particular, we construct spatial features within a complex representation model, then use the features as inputs to a simpler predictive model (like a generalized linear model). Our approach generates predictions with smaller bias and smaller variance than other spatial interpolation models such as bivariate splines in most situations. This method also enables us to generate rates in territories with no historical experience.