Journal of Risk Model Validation

This issue of The Journal of Risk Model Validation has a special section on model risk (beginning on page 77), so we have five papers rather than the usual four. I welcome innovations in format that offer added benefits to our readers. The format of this issue is three research papers followed by a special section of two papers. I am delighted to acknowledge the guest editors Christian Meyer and Peter Quell of DZ BANK AG, who very ably handled the demands of the special section. We hope they will do so again in future.

Our first paper, by Michael Jacobs Jr., is “Asset price bubbles and the quantification of credit risk capital with sensitivity analysis, empirical implementation and an application to stress testing”. It presents an analysis of the impact of asset price bubbles on standard credit risk measures. Jacobs performs a sensitivity analysis of the model parameters on the resulting credit risk measures as well as the changes in their relationship to the constant elasticity of variance (CEV) parameter. Readers are reminded that this framework links changes in volatility to changes in the price level; it is used by the author as a departure from fair value, which in itself is an interesting risk validation exercise, and captures some interesting bubble-like aspects.

Peter Thompson, Hayden Luo and Kevin Fergusson’s “The profit-and-loss attribution test”, the second paper in this issue, analyzes the failure probabilities of the profit-and-loss attribution (PLA) test, as defined in the final market risk standard published in January 2016 by the Basel Committee on Banking Supervision. The authors calculate both theoretical failure under standard distributional assumptions and the probabilities of failing the PLA test within different horizons. They also look at the steady-state proportion of desks that a bank might expect to maintain accreditation in order to use the internal model approach, assuming a minimum period of delay associated with the reaccreditation process subsequent to a desk failing the PLA test. Their analysis explains why the PLA test is likely to have a high failure probability, making it difficult to pass over a sustained period. In my opinion, this is a highly relevant study for practitioners.

Our third paper, “New historical bootstrap value-at-risk model” by Nikola Radivojevic, Zorana Sobat-Matic and Borjana B. Mirjanic, presents a new value-at-risk (VaR) model for the estimation of market risk in banks and other financial institutions. The model is called a new historical bootstrap VaR model because it shares the same theoretical basis as the historical simulation (HS) and bootstrap approaches. This paper aims to answer the following question: does incorporating the bootstrap method into the HS model contribute to improving the applicability of the HS approach with regard to meeting the backtesting rules of the Basel Accord? Based on an analysis of a subset of European capital markets, the authors claim that their new model performs better than the HS model. Such research, which uses data intensively, is very timely and may well lead to much future work in this area.

We now turn to the special section on model risk. Our first paper, and the issue’s fourth overall, is “A practical maturity assessment method for model risk management in banks” by L. van Biljon and L. J. Haasbroek. This proposes a qualitative method to assess the maturity of model risk management practices within banks. For some readers, a break from the remorseless flow of quant may be welcome. This method is aligned with relevant regulatory guidance and observed best practice. It gives banks a practical way to both determine their current maturity levels with respect to model risk management practices and define a targeted level of maturity. It also makes clear what aspects need to be remedied to progress from the current state to the targeted maturity state. We often forget that managing the investment horizon is another dimension of risk management.

Our final paper is Dennis E. Bennett’s “Governance and organizational requirements for effective model risk management”. This paper develops upon the foundation of model risk analytics and addresses the governance, organizational and human behavioral challenges of enterprise model risk management (MRM). Bennett proposes a comprehensive framework for model risk governance, organizational responsibilities and human behavior from the Risk Committee of the Board to MRM. The author also expands the definition of models to include decision support tools (DSTs), so that end-user computing, big data analytics and machine-learning DSTs are a part of model risk governance, organization, negotiation and control. Putting human behavior, which has the potential to include operational risk, together with big matters clearly opens up many new areas that may not have been addressed by more conventional approaches.

Steve Satchell
Trinity College, University of Cambridge

You need to sign in to use this feature. If you don’t have a account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here