Agent-based modelling for central counterparty clearing risk

Enhancing CCP Resilience, Recovery, and Resolution for financial institutions, exchanges, and policy makers

This paper proposes an agent-based-modelling approach to simulate an artificial CCP market environment. The model is intended to help market participants understand, analyse, and quantify the impacts of ongoing debates and proposals for the risk management framework at CCPs.

Banks must be able to effectively analyse and understand their bilateral relationships with CCPs. The COVID-19 crisis illustrates the imperative of doing this properly now so that they can be well prepared for the recovery and the next market shock. CCPs and policy makers also need to understand the system-wide dynamics resulting from the unique positions of market participants and their interactions, before proposing and implementing changes.

By using a simulated environment based on ABMs that can both realistically and easily represent the real-world, banks and other market participants can better discuss and test potential policies, processes and strategy changes. Decisions will therefore be based on a greater depth of insight into a wider representation of system dynamics. This is because ABMs more accurately model the emergent (and irrational) behaviour found in the real world.

With risking levels of central clearing after the 2008-09 financial crisis, there has been growing concern about the effectiveness of CCP risk management among market participants and policy makers. In 2019, a large group of banks and buy side firms discussed how to strengthen the risk management processes at CCPs through a joint publication - A Path Forward for CCP Resilience, Recovery and Resolution1. Their recommendations focused on addressing outstanding issues at CCPs through enhanced risk management and aligned interests among market participants.

In this report, we propose an agent-based-modelling approach to simulate an artificial CCP market environment, helping participants to understand, analyse, and quantify the effects of the proposed changes in the joint publication by the financial institutions. Deloitte’s ABM model is built jointly with Simudyne, a simulation software company, and Cloudera, an enterprise data cloud company. The combined technologies provide a highly scalable, cloud-enabled simulation solution.

1 The paper was originally published and backed at a firm-wide level in October 2019 by Allianz, BlackRock, Citi, Goldman Sachs, JPMorgan Chase & Co., Societe Generale, State Street, T. Rowe Price, and Vanguard. Support for the paper has recently been extended to include ABN AMRO Clearing, Barclays, Deutsche Bank, Commonwealth Bank of Australia, Franklin Templeton, Guardian Life, Ivy Investments, Nordea, TIAA and UBS as of March 2020.

Agent-based modelling for central counterparty clearing risk
Did you find this useful?