Optimizing the Compression Cycle: Algorithms for Multilateral Netting in OTC Derivatives Markets

Abstract

Concerns about the fragility of the financial system caused by the OTC derivatives market has encouraged the increased use of counterparty risk mitigation techniques including the use of market compression. In this process groups of market participants share position information via a third party company, TriOptima, who determine and propose a set of trades which will allow these parties to net out, i.e. "compress", their bilateral and multilateral exposures.

In this paper we introduce, analyse and compare a set of compression algorithms to see how successful they are at compression. Our main focus is the credit derivatives market, although the five methods considered here can also be used in other OTC derivative markets such as interest rate swaps. We compare all of these approaches using Monte-Carlo techniques and find a local pruning algorithm that performs as well as global approaches with the advantage that it avoids small positions and is very fast.

We believe that compression is a serious alternative counterparty risk mitigation technique to the use of CCPs that should be encouraged by regulators since it enables multilateral contract netting within the framework of an ISDA Credit Support Agreement thereby maintaining the added benefit of netting across all mutual ISDA derivative contracts.

Keywords: Counterparty, CDS, OTC derivatives, default, algorithm, CDS indices, Numerical methods

JEL Classification: G12, G13, G14, G15