Financial Engineering Workshops - Giuseppe Bonavolontà and Alper Hekimoglu

Title. Quantitative Model Validation in Structured Credit: Numerical Benchmarks and Challenger Approaches

Abstract

In this talk, presented from the perspective of practitioners, we discuss quantitative model validation as a core component of model risk management in structured credit. We illustrate this perspective through two concrete case studies.

The first part of the presentation addresses portfolio credit risk under multi-factor copula frameworks. Industry practice still relies heavily on Monte Carlo simulation for the estimation (contribution analysis) of tail risk measures such as VaR and Expected Shortfall. As a challenger methodology, we explore the Fourier–Cosine (COS) method of Fang - Oosterlee, originally developed for option pricing, and extend it to structured credit portfolios with guarantees and tranche structures (part of ongoing project with Prof. F. Fang and G. Maast at TU Delft).

The second part turns to hybrid callable bonds with credit and interest-rate risk modeled via correlated CIR processes. Here, validation requires reconciling two fundamentally different numerical paradigms: PDE-based valuation schemes, and simulation-based Longstaff–Schwarz regression.

Alper Hekimoglu.

Phd financial mathematics middle east technical university MSc Financial engineering Claremont Graduate University worked in ING market risk validation, rabobank and central bank of Turkey as quant economist in financial markets analysis division and banking expert in banking regulation and supervison agency of Turkey. Has published several papers in quantitative finance

Giuseppe Bonavolonta.

Head of Quantitative Model Validation unit (credit risk, ALM, PE, capital modeling and green finance modeling) holds a Ph.D. in Mathematics with expertise in the development, validation, regulatory framework and governance of financial models.