An Examination of European Firms’ Derivatives Usage: The Importance of Model Selection

• Date: 01 September 2017
• Author: James Ryan,Fergal O'Brien,Anthony Carroll
• Published date: 01 September 2017
• DOI: http://doi.org/10.1111/eufm.12115

An Examination of European Firms’

Derivatives Usage: The Importance of

Model Selection

Anthony Carroll, Fergal O’Brien and James Ryan

Kemmy Business School, University of Limerick, Limerick, Ireland

E-mails: anthony.g.carroll@ul.ie, fergal.g.obrien@ul.ie, james.ryan@ul.ie

Abstract

This paper investigates the determinants of foreign currency (FX) and interest rate

(IR) derivatives usage for European non-financial firms. We employ a Tobit model

and a two-part model which allows the determinants of the usage decision to differ

from the extent of usage decision. We find FX derivatives usage is motivated by

economies of scale and FX exposure, while IR derivatives usage is motivated by the

magnitude and nature of firms’debt. We also find that for IR derivatives the

determinants of the usage decision differ from the determinants of the extent of

usage decision.

Keywords: foreign exchange exposure, interest rate exposure, hedging policies

JEL classification:G32

1. Introduction

The 2007–2008 crisis in the capital markets, which wreaked havoc in the financial sector,

caused relatively few problems in the non-financial sector. Bartram et al. (2015) contend

that the reason for this is that economic risk (e.g., competition, price hikes in the factors

of production, etc.) accounts for the vast majority of non-financial firms’stock volatility

whereas their exposure to capital markets, i.e., market-based financial risks such as those

pertaining to interests rates and foreign currencies, accounts for only a fraction. This is

consistent with such firms managing their financial risk, which they can do by means of

adjustments to leverage, dividend and liquidity policies but also by hedging this risk

using financial derivatives. Moreover, the established theory suggests that corporate risk

The authors are grateful to the anonymous referees and John Doukas (the Editor) for their

very helpful comments, suggestions and support throughout the review process. We are

further grateful to the Kemmy Business School for its funding throughout this research

project. Anthony Carroll acknowledges the financial support received from the Irish

Research Council.

European Financial Management, Vol. 23, No. 4, 2017, 648–690

doi: 10.1111/eufm.12115

© 2017 John Wiley & Sons, Ltd.

management can not only mitigate risk but also increase firm value (Froot et al., 1993;

Leland, 1998). However, conclusive empirical evidence that corporate risk management,

and hedging with financial derivatives in particular, matters for firm value is mixed, with

some studies finding a positive relationship (Carter and Sinkey Jr, 1998; Disatnik et al.,

2014; MacKay and Moeller, 2007) but others finding no significant relationship (Bartram

et al., 2011; Guay and Kothari, 2003; Jin and Jorion, 2006). Chang et al. (2016) suggest

one reason for this inconsistency is that even sell-side analysts, despite their apparent

financial expertise, routinely misjudge the earnings implications of firms’derivatives

activities. What hope then for the average shareholder? All of the above therefore

necessitates a more granular investigation into what motivates firms to use derivatives,

the types of derivatives they use, and the level of that usage.

Previous studies have attempted to do so but many suffer from externally imposed

restrictions such as availability of suitable data. Many therefore limit the scope of their

investigations to certain categories of derivatives, or to particular industries and

geographic locations. Furthermore, a large variety of econometric models and

explanatory variables have been employed across these studies, making comparisons

between each difficult. Table 1 breaks down a selection of these studies by geographic

location of the samples selected, the scope of their investigations, and the models

employed.

With respect to scope, the third column illustrates how much of the previous research

focuses on the determinants of certain classifications of derivatives usage –foreign

currency (FX), interest rate (IR) or commodity price (CP) –whereas others examine

general, or aggregate, derivatives usage. Examining general derivatives usage

presupposes that the different categories of derivatives share common determinants.

However, if the determinants of FX, IR and CP derivatives usage differ from one another,

then they should be disaggregated and examined separately. Furthermore, comparisons

between different studies should be restricted to only those that examine the same

classification of derivatives.

The fourth column shows the range of models employed to examine firms’derivatives

usage. Studies that exclusively examine the usage decision employ either a logit or a

probit model. Generally, the requirement is that a firm that discloses derivatives usage,

even qualitatively, be classified as a derivatives user. Hence, firms with widely varying

magnitudes of derivatives usage are broadly categorised as users, which can introduce a

large amount of measurement error. Studies that examine the extent of firms’derivatives

usage, most commonly measured with gross notional amounts, use either a Tobit or a

two-part model, which differ in their assumptions of firms’decision processes: those that

assume that the binary derivatives usage and extent of usage are jointly decided in one

step employ the Tobit model, whereas those that assume firms decide on the extent of

usage only after they have decided to use derivatives employ a two-part or hurdle model.

Perhaps a reason for the apparent disparity in empirical findings is the lack of consensus

on which model is the most representative of firms’actual decision processes.

In addition to scope and model employment, studies also vary in their selection of

explanatory variables. Of the ten studies that examine the determinants of FX derivatives

usage in Table 1, only five control for FX exposure with some measure for foreign sales

(Allayannis and Ofek, 2001; Bartram et al., 2009; Elliott et al., 2003; G

eczy et al., 1997;

Muller and Verschoor, 2005). Of the six studies that examine the determinants of IR

derivatives usage, only one attempts to directly control for IR exposure by measuring

variable rate debt (Chernenko and Faulkender, 2012). Graham and Rogers (2002), who

© 2017 John Wiley & Sons, Ltd.

An Examination of European Firms’Derivatives Usage649

examine the determinants of general derivatives usage, include a proxy for interest rate

exposure in the sum of short-term and floating rate debt (as a percentage of total debt).

The majority of these studies have been motivated by the theories of optimal hedging,

which argue that a range of firm attributes, such as its growth opportunities (Froot et al.,

1993) or the probability that it will get into financial distress (Smith and Stulz, 1985),

determines derivatives usage. Consequently, many studies share a common subset of

explanatory variables. However, even within this subset, empirical findings have been

Table 1

Overview of derivatives literature

This table provides an overview of the literature that uses annual report disclosures to examine the

determinants of firms’derivatives usage. The column entitled ‘Scope’refers to what categories of

derivatives are examined. For example, ‘FX and IR’means that study examines the determinants of FX

and IR derivatives separately. ‘General’refers to the determinants of aggregate derivatives usage, or

FX, IR and CP derivatives usage, combined. The last column, entitled ‘Model’, refers to the model

employed to examine firms’derivatives usage.

Author(s)SampleScope Model

G

eczy et al. (1997)USFX Logit

Allayannis and Ofek (2001)USFXTwo-part

Elliott et al. (2003)USFX OLS

Visvanathan (1998)USIRLogit

Borokhovich et al. (2004) USIRTobit

Chernenko and Faulkender (2012)USIROLS

Tufano (1996)USCPTobit

Haushalter (2000)USCPTwo-part and Tobit

Dionne and Garand (2003)USCPTobit

Adam and Fernando (2006)USCPTwo-part

Mian (1996)USFX, IR and CPLogit

Howton and Perfect (1998)USFX and IRTobit

Gay and Nam (1998)USGeneralTobit

Fok et al. (1997)USGeneral Logit

Graham and Rogers (2002)USGeneralTobit

Knopf et al. (2002)USGeneral Tobit

Rogers (2002)USGeneralTobit

Lin and Smith (2007)USGeneralProbit

Nguyen and Faff (2002)AustraliaGeneralTobit

Nguyen and Faff (2003)AustraliaFX and IRTobit

Marsden and Prevost (2005)New ZealandGeneralTwo-part

Bartram et al. (2009)International (incl.

Europe)

FX, IR and CPProbit

Lievenbr€

uck and Schmid (2014)International (incl.

Europe)

FX, IRand CPProbit and Tobit

Lel (2012)International (incl.

Europe)

FX Tobit

Judge (2006)UKGeneralLogit

Muller and Verschoor (2005)Belgian, Dutch, German, UK FXTwo-part

© 2017 John Wiley & Sons, Ltd.

650 Anthony Carroll, Fergal O’Brien and James Ryan