Various Approximations of the Total Aggregate Loss Quantile Function with Application to Operational Risk
Posted: 15 Feb 2016 Last revised: 13 Nov 2016
Date Written: February 13, 2016
A compound Poisson distribution is a sum of independent and identically distributed random variables over a count variable that follows a Poisson distribution. Generally, its distribution is not tractable. However, it has many practical applications that require the estimation of the quantile function at a high percentile, e.g. 99.9th percentile. Without loss of generality, this paper focuses on its application to operational risk. We assume that the support of random variables is non-negative, discrete and finite. We investigate the mechanics of the empirical aggregate loss bootstrap distribution and suggest different approximations of its quantile function. Further, we study the impact of empirical moments and large losses on the empirical bootstrap capital at 99.9th confidence level.
Keywords: Loss sample, Frequency, Discrete Compound Distribution, Total aggregate loss, Quantile function, Analytic approximation, Extended Poisson, Extended Binomial, Cornish-Fisher, Panjer Recursion, Monte-Carlo simulation, Single Loss Approximation
JEL Classification: C00, C13, C15
Suggested Citation: Suggested Citation