Seminar by Hong Li

Pricing Catastrophe Bonds --- A Probabilistic Machine Learning Approach

This paper proposes a probabilistic machine learning method to price catastrophe bonds in the primary market. The proposed method combines machine-learning-based predictive models with conformal predictions, and is able to generate distribution-free probabilistic predictions for catastrophe bond prices. Using CAT bond data issued between January 1999 and March 2021, the proposed method is found to be more robust and yields more accurate predictions of the bond spreads than traditional methods such as linear regression. Furthermore, the proposed method generates more informative prediction intervals than learning regression and identify important nonlinear relationships between different risk factors and bond spreads, suggesting that linear regression could mis-estimate the bond spreads.

Overall, this paper demonstrates the potential of machine learning methods in improving the pricing of catastrophe bonds in the primary market.