Journal of Operational Risk

Welcome to the fourth issue of Volume 12 of The Journal of Operational Risk.

We have received a number of excellent papers on the advanced measurement approach/standardized measurement approach (AMA/SMA) debacle. While the regulatory path for operational risk capital remains uncertain, the industry is reacting by offering suggestions and showing the limitations of both approaches. In this issue, we feature two papers that discuss the AMA and SMA. One of our regular authors, Fabio Piacenza (with Claudia Belloni), discusses how the SMA could incorporate insurance, the only known financial operational risk hedging, into capital calculation. In another paper, Andrés Mora-Valencia provides a summary of the studies comparing both approaches. Over the next few issues we will have a number of papers on this subject, including some by regulators. Stay tuned.

The Journal of Operational Risk, as the leading publication in this area, aims to be at the forefront of these discussions. We welcome papers that shed light on them.

RESEARCH PAPERS

In our first paper, “Fast, accurate and straightforward extreme quantiles of compound loss distributions”, John Douglas (“J.D.”) Opdyke presents an easy-to-implement, fast and accurate method to approximate extreme quantiles of compound loss distributions (frequency C severity), which are commonly used in insurance and operational risk capital models. Opdyke uses an interpolated single-loss approximation (ISLA) method based on the widely used single-loss approximation (SLA) method, which was published by M. Degen in Volume 5, Issue 4 of The Journal of Operational Risk. In the author’s view, ISLA maintains two important advantages over SLA. First, ISLA correctly accounts for a discontinuity in SLA that can otherwise bias the approximation under conditions of both finite and infinite mean. Second, because it is based on a closed-form approximation, ISLA maintains the notable speed advantages of SLA over other methods requiring algorithmic looping (eg, fast Fourier transform or Panjer recursion). Speed is important when simulating many capital estimates, as is so often required in practice, and it is essential when extensive scenario simulations are needed (eg, in some power studies). The modified ISLA (MISLA) method also presented in this paper increases the range of application across the severity distributions most commonly used in these settings; it is tested against extensive Monte Carlo simulation, and it is arguably more straightforward to implement for the majority of AMA banks that are already using SLA.

“Standardized measurement approach extension to integrate insurance deduction into operational risk capital requirement” is this issue’s second paper. In it, Fabio Piacenza and Claudia Belloni claim that one of the main problems with the SMA is that it does not allow the inclusion of insurance coverage as capital requirement deduction. As a direct consequence, the SMA offers no incentives to invest in insurance coverage in order to keep the risk profile under control. The authors state that even the incentive to invest in other mitigation actions is reduced, since forward-looking components are not considered and it takes several years to significantly affect the SMA capital requirement through loss reduction. This paper describes a possible proposal to extend the SMA to include insurance coverage. The operational risk capital-at-risk (OpCaR) model – probably the same one used to calibrate the SMA – is a natural choice for integrating insurance coverage into the extended SMA.

FORUM SECTION

In Andrés Mora-Valencia’s “A note on the standard measurement approach versus the loss distribution approach–advanced measurement approach: the dawning of a new regulation”, a nonexhaustive review of the literature on operational risk quantification under a combination of the loss distribution approach model (the most commonly used of the AMA models) and extreme value theory is presented. The literature review points out that Bayesian inference has provided solutions to different problems when modeling operational data. The main comments prepared by the financial industry in response to the new proposal as well as two recently published papers that analyze the impact of SMA are also summarized. Finally, the discussion section proposes an alternative solution: an SLA model (taking into account several severity and frequency distributions) with an appropriate risk.

In the final paper of this issue, “Toward an efficient people-risk capital allocation for financial firms: evidence from US banks”, José Manuel Feria-Dominguez and Enrique Jimenez-Rodriguez claim that banks must allocate regulatory capital for covering their people-risk exposure. By using the Algo OpDataTM data set from US banks, and based on the loss distribution approach, they first estimate people-value- at-risk (people-VaR), assuming perfect correlation among people-risk categories but nonperfect dependence, for which the multivariate fast Fourier transformation is proposed. The diversified people-VaR is provided as a key indicator of an efficient capital allocation, and the traditional risk-adjusted return on capital measure is then readapted to evaluate the people-risk-adjusted performance.

Marcelo Cruz