Risk‐Sharing or Risk‐Taking? Counterparty Risk, Incentives, and Margins

• Author: FLORIAN HEIDER,MARIE HOEROVA,BRUNO BIAIS
• DOI: http://doi.org/10.1111/jofi.12396
• Published date: 01 August 2016
• Date: 01 August 2016

THE JOURNAL OF FINANCE •VOL. LXXI, NO. 4 •AUGUST 2016

Risk-Sharing or Risk-Taking? Counterparty Risk,

Incentives, and Margins

BRUNO BIAIS, FLORIAN HEIDER, and MARIE HOEROVA∗

ABSTRACT

Derivatives activity,motivated by risk-sharing, can breed risk-taking. Bad news about

the risk of an asset underlying a derivative increases protection sellers’ expected

liability and undermines their risk-prevention incentives. This limits risk-sharing,

creates endogenous counterparty risk, and can lead to contagion from news about the

hedged risk to the balance sheet of protection sellers. Margin calls after bad news

can improve protection sellers’ incentives and in turn enhance risk-sharing. Central

clearing can provide insurance against counterparty risk but must be designed to

preserve risk-prevention incentives.

DERIVATIVES ACTIVITY HAS GROWN strongly over the past 15 years. For example,

credit default swaps (CDS), bilateral over-the-counter contracts used to insure

credit risk, alone saw total notional amounts outstanding increase from around

$180 billion in 1998 to a peak of over $60 trillion by mid-2008 (Acharya et al.

(2012)). But the insurance provided by derivatives is effective only if counter-

parties can honor their contractual obligations. When Lehman Brothers filed

for bankruptcy protection in September 2008, for instance, it froze the positions

of more than 900,000 derivative contracts (about 5% of all derivative transac-

tions globally) in which Lehman Brothers was a party (Fleming and Sarkar

(2014)).

The above observations lead to several questions. First, what are the inter-

actions between counterparty risk and derivatives activity? Second, can risk-

sharing via derivatives lead perversely to risk-taking by financial institutions?

And third, how can derivatives activity be made more resilient to risk? This

paper seeks to address these questions. In particular, we explain how deriva-

tives positions affect risk-taking incentives, we show how margin deposits and

clearing arrangements can be designed to mitigate counterparty risk, and we

∗Biais is with the ToulouseSchool of Economics (CRM/CNRS IDEI). Heider and Hoerova are with

the European Central Bank and CEPR. We would like to thank the Acting Editor (Markus Brun-

nermeier), an Associate Editor, two anonymous referees, our discussants Ulf Axelson, Jonathan

Berk, Sugato Bhattacharyya, Bruce Carlin, Simon Gervais, Artashes Karapetyan, Christine Par-

lour, Alp Simsek, and Lauri Vilmi, as well as numerous seminar and conference participants for

their comments and suggestions. The views expressed do not necessarily reflect those of the Euro-

pean Central Bank or the Eurosystem. Biais gratefully acknowledges the support of the European

Research Council (Grant 295484-TAP). We have read the Journal of Finance’s disclosure policy

and have no conflicts of interest to disclose.

DOI: 10.1111/jofi.12396

1669

1670 The Journal of Finance R

provide new empirical predictions about the extent of derivatives activity in a

given financial environment and the default risk of institutions selling protec-

tion through derivatives.

Our model features risk-averse protection buyers who want to insure against

a common exposure to risk (any idiosyncratic component of risk can be diver-

sified away among protection buyers themselves). To insure the common risk,

protection buyers contact risk-neutral protection sellers whose assets can be

risky, but who are not directly exposed to the risk that buyers want to insure.

Because of limited liability, protection sellers can make insurance payments

only if their assets are sufficiently valuable. The value of a protection seller’s

assets is affected by her actions. Specifically, protection sellers can maintain

the value of their assets, and in turn prevent downside risk, by exerting costly

effort. For example, protection sellers can put in place strong internal risk

management systems (see Ellul and Yerramilli (2013)) and carefully scruti-

nize the quality of potential investments before undertaking them. However,

protection sellers may avoid costly risk-prevention effort by choosing weaker

internal risk controls or by relying on external credit ratings and simple

backward-looking measures of risk. The failure of protection sellers to exert

risk-prevention effort (what we call “risk-taking” ) leads to counterparty risk

for protection buyers. Since financial institutions’ balance sheets and activities

are opaque and complex, a lack of risk-prevention effort is difficult for outsiders

to detect. This creates a moral hazard problem for protection sellers, the key

friction in our model.

Our model builds on two important characteristics of derivatives activity.

First, during the life of a derivative contract, new information about the value of

an underlying asset becomes available. Such news affects the expected payoffs

of the contracting parties: it makes the derivative position an asset for one

party and a liability for the other. Second, derivative exposures, and hence the

associated potential liabilities, can be large. According to the Quarterly Report

on Bank Derivatives Activities by the Office of the Comptroller of the Currency,

total credit exposure from derivatives reached more than $1.5 trillion in 2008.1

The total credit exposure of the top five financial institutions was 2 to 10 times

larger than their risk-based capital.

A key insight of our analysis is that a large derivative exposure undermines

a protection seller’s incentives to exert risk-prevention effort when news makes

the derivative position an expected liability for her. In that case, she bears the

full cost of risk-prevention effort while part of this effort’s benefit accrues to

her counterparty in the form of payments from the derivative contract. This

mechanism is reminiscent of debt overhang (Myers (1977)) but there is an

important difference. In our analysis, the liability arises endogenously in the

1Total credit exposure is the sum total of net current credit exposure (NCCE) and potential

future exposure (PFE). NCCE is the gross positive fair value of derivatives contracts less the

dollar amount of netting benefits. PFE is an estimate of what the current credit exposure could be

over time, based upon a supervisory formula in the risk-based capital rules.