Monte Carlo Estimation of CoVaR

Abstract

CoVaR is an important measure of financial systemic risk due to its ability to capture tail dependence between the losses of different portfolios and its capacity to predict financial crises. Estimating CoVaR is challenging because its definition involves a zero-probability event, which is unobservable in the data. The existing model-based methods address this issue by assuming simplified structural models, which introduce biases that are difficult to eliminate. In “Monte Carlo Estimation of CoVaR,” Huang, Lin, and Hong propose using Monte Carlo methods to estimate CoVaR, leveraging the modeling flexibility of Monte Carlo simulation. Specifically, they introduce a batching estimator applicable to a wide range of financial models and prove that its best rate of convergence is [Formula: see text], where n is the sample size. Under the widely used delta-gamma approximation model, they further introduce an importance sampling–inspired estimator and prove that its best rate of convergence can be improved to [Formula: see text].