A nonparametric urn-based approach to interacting failing systems with an application to credit risk modeling

TL;DR

This paper introduces a nonparametric urn-based recursive model for interacting failing systems, such as firms and bonds, to estimate default probabilities and failure distributions, with an application to credit risk assessment.

Contribution

It presents a novel nonparametric urn-based recursive framework for modeling interacting failing systems, extending credit risk modeling capabilities.

Findings

Effective assessment of obligor default probabilities

Joint failure probability estimation for risk portfolios

Application demonstrates practical utility in credit risk analysis

Abstract

In this paper we propose a new nonparametric approach to interacting failing systems (FS), that is systems whose probability of failure is not negligible in a fixed time horizon, a typical example being firms and financial bonds. The main purpose when studying a FS is to calculate the probability of default and the distribution of the number of failures that may occur during the observation period. A model used to study a failing system is defined default model. In particular, we present a general recursive model constructed by the means of inter- acting urns. After introducing the theoretical model and its properties we show a first application to credit risk modeling, showing how to assess the idiosyncratic probability of default of an obligor and the joint probability of failure of a set of obligors in a portfolio of risks, that are divided into reliability classes.