2008. An efficient simulation procedure for point estimation of expected shortfall. Working paper 08-03

The authors thank Hai Lan for providing assistance with computer code and experiments. Measuring a portfolio’s risk accurately can be computationally expensive when it is necessary to revalue the portfolio in many scenarios using Monte Carlo simulation. We present and test a com-putationally efficient simulation procedure for point estimation of expected shortfall, a risk measure closely related to conditional value at risk and tail conditional expectation. The procedure applies tools from the literature on ranking and selection by simulation to allocate more computational resources to estimation of the largest losses, which are those that affect expected shortfall. In our experiments, given a fixed computational budget, our procedure estimates expected shortfall with a much lower mean squared error than a standard simulation procedure. It also produces a point estimator much more precise than the interval estimator of Lan et al. (2008).