Efficient Machine Learning Methods for Risk Management of Large Variable Annuity Portfolios
Variable annuity (VA) embedded guarantees have rapidly grown in popularityaround the world in recent years. Valuation of VAs has been studiedextensively in past decades. However, most of these studies focus on asingle contract. These methods cannot be extended to valuate a largevariable annuity portfolio due to the computational complexity. In thispaper, we propose an efficient moment matching machine learning method tocompute the annual dollar deltas, VaRs and CVaRs for a large variableannuity portfolio whose contracts are over a period of 25 years. There aretwo stages for our method. First, we select a small number of contracts andpropose a moment matching Monte Carlo method based on the Johnson curve,rather than the well known nested simulations, to compute the annual dollardeltas, VaRs and CVaRs for each selected contract. Then, these computedresults are used as a training set for well known machine learning methods,such as regression tree, neural network and so on. Afterwards, the annualdollar deltas, VaRs and CVaRs for the entire portfolio are predicted throughthe trained machine learning method. Compared to other existing methods, ourmethod is very efficient and accurate, especially for the first 10 yearsfrom the initial time. Finally, our test results support our claims.