A stochastic switching control model arising in general OTC contracts with contingent CSA in presence of CVA, collateral and funding

TL;DR

This paper introduces a stochastic switching control model for OTC contracts with a contingent CSA, optimizing collateralization strategies to minimize counterparty risk and funding costs, using advanced stochastic control and numerical methods.

Contribution

It develops a novel stochastic switching control framework for contingent collateral agreements in OTC contracts, incorporating CVA, collateral, and funding costs, with theoretical and numerical solutions.

Findings

The model demonstrates potential cost savings over fixed collateralization.

Numerical algorithms enable practical computation of optimal switching strategies.

The approach integrates CVA, collateral, and funding costs into a unified control framework.

Abstract

The present work studies and analyzes general defaultable OTC contract in presence of a contingent CSA, which is a theoretical counterparty risk mitigation mechanism of switching type that allows the counterparty of a general OTC contract to switch from zero to full/perfect collateralization and switch back whenever she wants until contract maturity paying some switching costs and taking into account the running costs that emerge over time. The motivation and the underlying economic idea is to show that the current full/partial collateralization mechanisms defined within contracts' CSA - and now imposed by the banking supervision authorities - are "suboptimal" and less economic than the contingent one that allows to optimally take in account all the relevant driver namely the expected costs of counterparty default losses - represented by the (bilateral) CVA - and the expected collateral and funding costs. In this perspective, we tackle the problem from the risk management and optimal design point of view solving - under some working assumptions - the derived stochastic switching control model via Snell envelope technique and important results of the theory of the backward stochastic differential equations with reflection (RBSDE). We have also studied the numerical solution providing an algorithm procedure for the value function computation based on an iterative optimal stopping approach.