Heterogeneity and netting efficiency under central clearing: A stochastic network analysis

• Author: Injun Hwang,Baeho Kim
• Date: 01 February 2020
• DOI: http://doi.org/10.1002/fut.22059
• Published date: 01 February 2020

J Futures Markets. 2020;40:192–208.wileyonlinelibrary.com/journal/fut192

|

© 2019 Wiley Periodicals, Inc.

Received: 20 August 2019

|

Accepted: 27 August 2019

DOI: 10.1002/fut.22059

RESEARCH ARTICLE

Heterogeneity and netting efficiency under central

clearing: A stochastic network analysis

Injun Hwang

|

Baeho Kim

Korea University Business School, Seoul,

Republic of Korea

Correspondence

Baeho Kim, Korea University Business

School, Anam‐dong, Sungbuk‐Gu, Seoul

02841, Republic of Korea.

Email: baehokim@korea.ac.kr

Abstract

This paper examines the effect of heterogeneity in clearing members’exposure

management practices under central clearing. Our network model specifies the

dynamics of prenetted interbank exposures to shape interdependent exposure

distributions beyond normality. Employing over‐the‐counter derivatives market

data from the U.S. Office of the Comptroller of the Currency, our simulation

results indicate that heterogeneity in bank‐to‐bank exposure dynamics is

systemically desirable, while the entire system benefits more from the central

clearing in more homogeneous environments. Policymakers should incentivize

individual clearing members to enhance resiliency and stability in counterparty

exposure management to maximize netting efficiency under central clearing.

KEYWORDS

central clearing, exposure distribution, heterogeneity, netting efficiency, simulation, stochastic

network model

JEL CLASSIFICATION

C15; C46; G17; G21

1

|

INTRODUCTION

The 2007–2008 global financial crisis underscored the need for structural changes in over‐the‐counter (OTC) derivatives

markets to mitigate systemic risk. One of the reforms’major dimensions is a mandatory central clearing of standardized

OTC derivatives contracts. Under central clearing, central counterparties (CCPs) offset bilateral obligations for a single

net position between participants at the center of the system. The CCPs, in turn, provide a multilateral netting channel

to (a) reduce the expected exposure in the entire financial system and (b) wedge a bulkhead in OTC markets by

isolating individual entities from the systemic propagation of counterparty credit risk; see Duffie, Li, and Lubke (2010),

Jones and Pérignon (2013), and Pirrong (2011) for extensive review on central clearing counterparty, and Bae, Karolyi,

and Stulz (2003), Elliott, Golub, and Jackson (2014) and Glasserman and Young (2015) for propagation mechanism of

financial contagion.

1

However, as illustrated by Duffie and Zhu (2011), a central clearing scheme that is limited to parts

of asset classes may deprive clearing members of bilateral netting opportunities across different asset classes; that is,

introducing CCPs does not always reduce the total expected counterparty exposure. This is because only the payments

implied by the novated contracts are netted under central clearing, otherwise, the remaining payments could be netted

with other contracts of different asset classes. Cont and Kokholm (2014) extend the work of Duffie and Zhu (2011) by

1

Title VII of the Dodd‐Frank Wall Street Reform, the Basel Committee on Banking Supervision’s document on postcrisis reform BCBS (2017), the 2010 Consumer Protection Act, the G20’s 2009

movement, and the European Market Infrastructure Regulation (EMIR) are well‐known examples; see BCBS (2019b), BCBS (2017), BCBS (2019a), BCBS‐IOSCO (2015), and European Union (2012).

addressing the concept of heterogeneity in asset classes in terms of riskiness as well as the intermarket exposure

correlation.

2

In this study, we examine the system‐wide benefit of central clearing by gauging the reduction in total expected

counterparty exposure through introducing a CCP in the OTC derivatives markets. Our networked exposure model

specifies the dynamics of prenetted interbank exposure as a joint stochastic process that shapes interdependent bank‐to‐

bank exposure distributions beyond normality by extending specifications in the existing literature; for example, see

Duffie and Zhu (2011) and Cont and Kokholm (2014) among others. In reality, asymmetric dominance is often observed

between two counterparties, as the larger or more creditworthy bank tends to be more exposed to its counterparty, and

vice versa. Consequently, taking postnetted exposures as a model primitive may mislead the efficiency analysis in the

presence of heterogeneous exposure management practices. To circumvent these drawbacks, we manifest the reality by

specifying our model for the prenetted exposures per each of the asset classes as stochastic processes inducing skewed

and fat‐tailed exposure distributions at the end of risk horizons. The advantage of our modeling approach resides in its

flexibility in revealing higher‐order moments and thereby enabling holistic analyses of systemic risk based on more

sensible characteristics of interbank exposure distributions in the network.

3

Employing OTC derivatives market data provided by the U.S. Office of the Comptroller of the Currency, from a

supervisory point of view, we explore how heterogeneity in bank‐to‐bank exposure management practices and exposure

size affect netting efficiency under central clearing. In addition, by perturbing the prenetted exposure model

parameters, we delve into the effect of bank‐specific resiliency and stability in the management of interbank exposure

on multilateral netting efficiency. Regardless of clearing schemes, we find that heterogeneous interbank exposures are

systemically beneficial to mitigate the amount of potential losses in the OTC derivatives market in a time of stress. This

is because the strong positive correlation in exposure distributions coincides with bank tendencies to engage in more

homogeneous asset management practices that may make their exposure networks more systemically vulnerable

(Acharya, 2009). Most importantly, our findings indicate that the systemic benefit of central clearing becomes more

pronounced, as the comovement between individual exposures becomes stronger. In other words, the multilateral

netting benefit under central clearing outweighs the bilateral reduction of expected exposures within an environment of

systemically homogeneous exposure dynamics between clearing members.

We further find a negative relationship between the CCP benefit and the dispersion in clearing members’size,

measured by their notional outstanding OTC derivatives. We can derive a policy‐oriented implication from this result.

From the perspective of system‐wide netting efficiency, regulations are supposed to prevent large clearing members

from taking new positions that increase their exposures. Our proposed approach extends the scope of the systemic

implication of heterogeneity between clearing members as illustrated by Choi (2014) in that homogeneous management

of interbank exposure can improve aggregate netting efficiency under central clearing.

Moreover, our simulation results indicate that the CCP benefit is significantly increasing in resiliency and stability

implied by the parameters of prenetted exposure processes. Specifically, the CCP benefit is unduly responsive to

changes in bank‐specific speeds of mean‐reversion as well as volatilities of exposure processes. Our findings

demonstrate that regulators should allocate more burden of loss mutualization to clearing members with more

positions on inelastic and volatile assets.

Prior research such as Duffie and Zhu (2011) and Cont and Kokholm (2014) has viewed greater netting efficiency as

a benefit of central clearing. Our study inherits the same point of view, while our model specification employs a

prenetted version of interbank exposure dynamics in a stochastic network. By doing so, we provide a unique insight in

bank‐specific and system‐wide dimensions under the unified model framework. Moreover, Garratt and Zimmerman

(2018) explore the trade‐off between the mean and variance of net exposure under the generalized network structure,

which does not require a full specification of interconnectedness between participants. They show that the set of

multilateral networks, in which less net exposure is totalized, provides a smaller variance of net exposures. Loon and

Zhong (2014) examine how the CCP can mitigate counterparty risk between clearing members based on an event study

of the credit default swap (CDS) market. They find that the prudent role of the CCP increases settlement CDS spreads,

leading to the reduction of counterparty risk under central clearing. Pereira da Silva, Vieira, and Vieira (2018)

2

Under the assumption of homogeneity regarding asset class riskiness, improving system‐wide netting efficiency via CCP requires an unrealistically large number of central clearing members (461 in

Duffie and Zhu, 2011). Meanwhile, under the heterogeneity assumption and using the fat‐tailed exposure distribution model generates a more down‐to‐earth number of clearing members (14 in Cont

and Kokholm, 2014) for the CCP to improve the netting efficiency at the system level.

3

The benefit of a prudent exposure model goes beyond netting efficiency. Duffie, Scheicher, and Vuillemey (2015) document the relationship between central clearing schemes and system‐wide

collateral demand with party‐to‐party bilateral credit default swap exposure data. As the netting efficiency affects the initial margin requirement, the trade‐off between bi‐and multi‐lateral netting

plays a key role in changing collateral demand with the addition of CCPs. This finding highlights the importance of a sensible and realistic exposure model.

HWANG AND KIM

|

193