Improving Computational Efficiency in Catastrophe Bond Valuation: Numerical Integration through the Multi-level Monte Carlo and

Direct modelling of the aggregate loss distributions for catastrophe bond pricing is complicated due to an often-intractable convolution function. Alternative approaches, including simulation and other approximations, have therefore been proposed. This study analyses the application of two techniques, the Multilevel Monte Carlo and the Expectation Maximization (EM) algorithm, to catastrophe loss modelling. Each technique’s computational efficiency and applicability for different bond payoff functions will be analysed. Further tests will then compare these results to those from previously applied techniques like the Quasi-Monte Carlo and other mixed approximation methods to establish model efficiency.