Abstract:We consider the problem of estimating the probability of a large loss from a financial portfolio, where the future loss is expressed as a conditional expectation. Since the conditional expectation is intractable in most cases, one may resort to nested simulation. To reduce the complexity of nested simulation, we present an improved multilevel Monte Carlo (MLMC) method by using quasi-Monte Carlo (QMC) to estimate the portfolio loss in each financial scenario generated via Monte Carlo. We prove that using QMC can accelerate the convergence rates in both the crude nested simulation and the multilevel nested simulation. Under certain conditions, the complexity of the proposed MLMC method can be reduced to O(ε−2(logε)2) . On the other hand, we find that using QMC in MLMC encounters a high-kurtosis phenomenon due to the existence of indicator functions. To remedy this, we propose a smoothed method which uses logistic sigmoid functions to approximate indicator functions. Numerical results show that the optimal MLMC complexity O(ε−2) is almost attained even in moderate high dimensions.

Conditional Monte Carlo Estimation of Quantile Sensitivities

Estimating Quantile Sensitivities

On the Asymptotic Analysis of Quantile Sensitivity Estimation by Monte Carlo Simulation.

Estimating Sensitivities of Portfolio Credit Risk Using Monte Carlo

Sensitivity estimation of conditional value at risk using randomized quasi-Monte Carlo

Pathwise Estimation of Probability Sensitivities Through Terminating or Steady-State Simulations

Conditional expectation method for estimating parameter sensitivity of unreliability

Estimating Confidence Intervals and Regions for Quantiles by Monte Carlo Simulation

Simulating Sensitivities of Conditional Value at Risk

Sensitivity Analysis of Portfolio Credit Derivatives by Conditional Monte Carlo Simulation

Monte Carlo Estimation of Value-at-risk, Conditional Value-at-risk and Their Sensitivities

Quasi-Monte Carlo Simulation for American Option Sensitivities

Confidence Intervals and Regions for Quantiles Using Conditional Monte Carlo and Generalized Likelihood Ratios.

Differential Quantile-Based Sensitivity in Discontinuous Models

Nonparametric Inference for Sensitivity of Haezendonck–Goovaerts Risk Measure

Convergence analysis of quasi-Monte Carlo sampling for quantile and expected shortfall

Conditional Empirical Likelihood for Quantile Regression Models

Estimating Quantile Sensitivity for Financial Models with Correlations and Jumps

Sensitivity Analysis of CoVaR.

Variance Reduction for Generalized Likelihood Ratio Method in Quantile Sensitivity Estimation

Efficient risk estimation via nested multilevel quasi-Monte Carlo simulation