Quantifying Credit Portfolio sensitivity to asset correlations with interpretable generative neural networks

TL;DR

This paper introduces a novel method using Variational Autoencoders to generate interpretable correlation matrices, enabling better understanding of credit portfolio sensitivity to asset correlation changes.

Contribution

It replaces GANs with VAEs for correlation matrix generation, providing more interpretability and insights into factors affecting credit portfolio risk.

Findings

VAE latent space effectively captures key diversification factors

The approach improves understanding of portfolio sensitivity to correlation shifts

Generated correlation matrices resemble empirical data

Abstract

In this research, we propose a novel approach for the quantification of credit portfolio Value-at-Risk (VaR) sensitivity to asset correlations with the use of synthetic financial correlation matrices generated with deep learning models. In previous work Generative Adversarial Networks (GANs) were employed to demonstrate the generation of plausible correlation matrices, that capture the essential characteristics observed in empirical correlation matrices estimated on asset returns. Instead of GANs, we employ Variational Autoencoders (VAE) to achieve a more interpretable latent space representation. Through our analysis, we reveal that the VAE latent space can be a useful tool to capture the crucial factors impacting portfolio diversification, particularly in relation to credit portfolio sensitivity to asset correlations changes.