Credit Exposure in the Presence of Initial Margin

20 Pages Posted: 8 Jul 2016 Last revised: 9 May 2017

Leif B. G. Andersen

Bank of America Merrill Lynch

Michael Pykhtin

Board of Governors of the Federal Reserve System

Alexander Sokol

CompatibL

Date Written: July 22, 2016

Abstract

We leverage the new framework for collateralized exposure modelling in Andersen, Pykhtin, and Sokol (2016) to analyze credit risk on positions collateralized with both variation and initial margin. Special attention is paid to the dynamic BCBS-IOSCO uncleared margin rules soon to be mandated for bilateral inter-dealer trading in OTC derivatives markets. While these rules set initial margin at a 99th 2-week percentile level and aim to all but eliminate portfolio close-out risk, we demonstrate that trade flow effects can result in exposures being reduced significantly less than expected. We supplement our analysis with several practical schemes for estimating IM on a simulation path, and for improving the speed and stability of exposure simulation. We also briefly discuss potential ways to adjust the margin framework to more effectively deal with exposures arising from trade flow events.

Keywords: margin period of risk, initial margin, credit exposure, uncleared margin rules, kernel regression

JEL Classification: G10, G13, G20, G33, C15

Suggested Citation

Andersen, Leif B. G. and Pykhtin, Michael and Sokol, Alexander, Credit Exposure in the Presence of Initial Margin (July 22, 2016). Available at SSRN: https://ssrn.com/abstract=2806156

Leif B.G. Andersen (Contact Author)

Bank of America Merrill Lynch ( email )

One Bryant Park
New York, NY 10036
United States
646-855-1835 (Phone)

Michael Pykhtin

Board of Governors of the Federal Reserve System ( email )

20th Street and Constitution Avenue NW
Washington, DC 20551
United States

HOME PAGE: http://www.federalreserve.gov/econresdata/michael-v-pykhtin.htm

Alexander Sokol

CompatibL ( email )

100 Overlook Center
Second Floor
Princeton, NJ 08540
United States

HOME PAGE: http://www.compatibl.com

Paper statistics

Downloads
470
Rank
47,771
Abstract Views
1,224