You are on page 1of 18

UK Multinationals Eective Use of Financial

Currency-Hedge Techniques: Estimating and


Explaining Foreign Exchange Exposure Using
Bilateral Exchange Rates
Stephen D. Makar and Stephen P. Human
College of Business Administration, University of Wisconsin Oshkosh, 800 Algoma Blvd.,
Oshkosh, WI 54901-8676, USA
e-mail: makar@uwosh.edu
Abstract
Using a unique dataset of recently available accounting disclosures, this study examines
the relationship between UK multinationals stock returns and changes in the principal
exchange rate to which each rm is most exposed. We nd more rms with signicant
foreign exchange exposure estimates using this rm-specic principal currency data,
compared with those exposure estimates using the broad exchange rate index data
prevalent in prior studies. The cross-sectional variations in such principal-currency
exposure estimates are explained in relation to the nancial currency-hedge techniques
that each rm specically identies as being used to manage its currency risk. In
particular, we provide evidence that rms eectively use foreign currency derivatives
and foreign-denominated debt to reduce the currency risk associated with the bilateral
exchange rate to which they are most exposed. This study is important to both the
academic and the practitioner communities because it represents the rst use of publicly
available UK disclosures to improve the estimation and explanation of foreign exchange
exposure.
1. Introduction
This study takes advantage of recently available accounting disclosures
to examine the exposure of UK multinationals to changing foreign
currency exchange rates, and their use of nancial hedges to manage
such currency risk. Although economic theory suggests that rm value is
related to contemporaneous exchange rate changes, there is a puzzling
lack of empirical support for market-based estimates of foreign exchange
exposure (Fa and Marshall, 2005). This gap in the foreign exchange
exposure literature has been attributed to three complementary explana-
tions (Dewenter et al., 2005). First, investors initially may lack informa-
tion sucient to understand a rms foreign exchange exposure, resulting
in a delayed market response to exchange rate changes (e.g., Bartov and
Bodnar, 1994). Second, rms may manage their currency risk eectively,
using foreign currency derivatives and other hedge techniques (e.g.,
Journal of International Financial Management and Accounting 19:3 2008
r 2008 Blackwell Publishing Ltd., 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
Allayannis and Ofek, 2001). Finally, there are methodological challenges
in estimating and explaining foreign exchange exposure, including the
identication of exchange rates to which the rm is exposed (e.g., Ihrig,
2001).
Using a rm-specic approach, we address each of these three
complementary explanations, and thus contribute to the foreign
exchange exposure literature. First, we estimate each rms foreign
exchange exposure in relation to the principal currency to which it is
exposed. In contrast, most prior studies have relied on either a broad
exchange rate index or a rm-specic exchange rate index based on the
geographic location of subsidiaries to estimate foreign exchange expo-
sure (e.g., Fraser and Pantzalis, 2004). Second, we explain the cross-
sectional variations in our principal-currency exposure estimates in
relation to hedge techniques that each rm specically identies as being
used to manage such risk, including foreign currency derivatives and
foreign-denominated debt. Previous research provides evidence consis-
tent with US rms using both of these nancial hedge techniques to
manage foreign exchange exposure (e.g., Allayannis and Ofek, 2001). By
using rm-specic hedge data to explain foreign exchange exposure
estimates particular to each rms principal currency, we respond to
prior studies call for rened measures of exposure and hedging variables
(e.g., Hagelin and Pramborg, 2004). Finally, we extend the return
horizon to accommodate any delayed market response to exchange
rate changes, as suggested by recent methodological papers (e.g., Bodnar
and Wong, 2003).
Following this rm-specic approach to estimating and explaining
foreign exchange exposure, we document UK multinationals eective
use of nancial currency-hedge techniques in the 19992002 sample
period. In particular, rms reporting the bilateral exchange rate to which
they are most exposed reduce such principal currency risk using foreign
currency derivatives and foreign-denominated debt. In addition, we nd
more rms with signicant exposure when our market-based analyses
incorporate this principal currency data versus using the exchange rate
index data prevalent in prior studies. This latter improvement in
estimating foreign exchange exposure is robust to the return horizon.
Likewise, our evidence of eective nancial currency hedging is not
sensitive to the rms size, its percentage of foreign sales, or its use of
non-nancial hedges. This study is important because it represents the
rst use of publicly available UK accounting disclosures to improve
the estimation and explanation of foreign exchange exposure. With
220 Stephen D. Makar and Stephen P. Human
r 2008 Blackwell Publishing Ltd.
the United Kingdom not joining the European Monetary Union, the
post-euro evidence of eective nancial-currency hedging for the 1999
2002 period is particularly compelling.
2. Background
In their seminal study, Adler and Dumas (1984) develop an approach to
measuring foreign exchange exposure based on the regression of stock
returns on the changes in exchange rates. They argue that currency risk is
similar to market risk, and therefore such foreign exchange exposure
estimates parallel the b estimates in the traditional market model.
1
Subsequent studies have relied on the Adler and Dumas approach at
the rm level, with limited success. For example, studies using a broad
exchange rate index for US multinationals report that the percentage of
rms exhibiting signicant exposure estimates ranges from 5% in Jorion
(1990) to 15% in Choi and Prasad (1995).
2
Moving beyond US studies using the Adler and Dumas (1984)
approach, Dominguez and Tesar (2006) nd that signicant exposures
to changes in broad exchange rate indices range from 5% of Chilean
rms to 26% of Japanese rms, with 11% of UK rms exhibiting
signicant market-based exposure estimates. Similarly, He and Ng
(1998) report that 25% of their sample of Japanese multinationals is
exposed to contemporaneous changes in a broad exchange rate index.
Doidge et al. (2006) also use trade-weighted exchange rates in measuring
the economic importance of currency exposure in 18 countries based on a
portfolio approach, while Bartram and Karolyi (2006) focus their multi-
country trade-weighted exchange rate analyses on the 1999 euro launch.
Employing an event study approach, Bartram and Karolyi report
currency exposure changes around the euro introduction that are
statistically and economically small. In contrast to these studies of
contemporaneous foreign exchange exposure, other researchers report
a lagged response of stock prices to exchange rate changes (e.g., Amihud,
1994; Bartov and Bodnar, 1994; Donnelly and Sheehy, 1996), and
attribute such results, in part, to the lack of timely information
concerning hedging activities. In light of this explanation for inconclusive
capital market evidence of foreign exchange exposure, Bodnar and Wong
(2003) recommend lengthening the return horizon to increase the
percentage of signicant exposure estimates.
As introduced in Section 1, the eective hedging of currency risk is a
second explanation for the limited success researchers have had in
UK Multinationals Financial Currency-Hedge Techniques 221
r 2008 Blackwell Publishing Ltd.
documenting signicant foreign exchange exposures. Indeed, Allayannis
and Ofek (2001) nd that the use of foreign currency derivatives
signicantly reduces market-based estimates of exchange rate exposure
for their sample of S&P 500 rms. In addition, they provide evidence
suggesting that US rms also use foreign-denominated debt to hedge
currency risk. In particular, Human and Makar (2004) report that US
foreign-denominated debt issuers use foreign currency derivatives to
hedge short-term risk eectively. Hagelin and Pramborg (2004) use a
nancial-hedge indicator variable (equal to 1 if either currency deriva-
tives or foreign-denominated debt is used) based on their questionnaire
data in documenting eective hedging of Swedish multinationals
exposure to changes in a broad exchange rate index. Similarly, DeJong
et al. (2006) rely on questionnaire data to provide evidence that Dutch
rms eectively use foreign-denominated debt and other on-balance
sheet hedges of rm-specic currency risk during the 19941998 pre-
euro period. Beyond such nancial currency-hedge techniques, other
studies have examined the relationship between non-nancial hedges and
estimates of exposure using a broad exchange rate index. Pantzalis et al.
(2001), for example, report that operational hedges are important
determinants of market-based exposure estimates for US multinationals,
while Kim et al. (2006) nd that both operational and nancial hedges
are associated with reductions in currency risk.
A third explanation for the gap in foreign exchange exposure evidence
is the diculty in identifying exchange rates to which the rm is exposed.
As reviewed above, previous studies using a broad exchange rate index
have found evidence of a relationship between rm value and exchange
rates, but only for a small portion of their samples (e.g., Jorion, 1990;
Choi and Prasad, 1995; He and Ng, 1998; Dominguez and Tesar, 2006).
Other recent studies have introduced rm-specic exchange rate indices
based on the geographic location of foreign subsidiaries. Ihrig (2001), for
example, reports that the number of signicant exposures using this
approach rises from 10% using a broad index to 16% of the sample of
US rms. Muller and Verschoor (2006) follow Ihrigs approach and
provide evidence of US rms asymmetric exposure to changes in region-
specic trade-weighted exchange rates. Similarly, Fraser and Pantzalis
(2004) construct an exchange rate index based on each rms geographic
operational network, and nd the percentage of signicant exposures
increases from 5.5% using a common index to 8.7% with the rm-
specic index. However, the authors also report that 12.6% of their US
sample exhibits signicant exposure to a more comprehensive common
222 Stephen D. Makar and Stephen P. Human
r 2008 Blackwell Publishing Ltd.
index; that is, when more currencies are included in a common index,
there is an increase in the number of rms with signicant exposures. In
conclusion, Fraser and Pantzalis emphasize that the selection of the
exchange rate variable is critical to the estimation and explanation of
foreign exchange exposure.
As an alternative rm-specic approach, prior research has used
particular bilateral exchange rates for samples with identiable currency
risk. For example, Glaum et al. (2000) focus on German multinationals
exposure to the US dollar, but caution that their daily market-based
exposure estimates exhibit notable variations over time and across rms.
Accordingly, the authors recommend that future researchers explore
such variations, including the impact of hedging activities. Williamson
(2001) focuses on US and Japanese automotive rms exposure to
changes in each countrys currency and in the German Mark. Despite
examining rms in an industry with known exposure, Williamson reports
a signicant rm value/exchange rate relation for only a subset of rms
and exchange rates. Williamson attributes such cross-sectional variation
in exposure, in part, to each rms eective hedging practices. Reviewing
the capital market evidence of bilateral exchange rate exposure, Dewen-
ter et al. (2005) note that these studies tend to report stronger evidence of
short-term exposure than those studies using a broad exchange rate
index.
In summary, the lack of empirical evidence of a relation between rm
value and exchange rate changes has been attributed, in part, to the
diculty in identifying rm-specic bilateral rates. Moreover, there is
evidence that such currency risk may be eectively hedged, and may
increase over longer return horizons. As introduced in Section 1, we are
able to identify the principal bilateral rate to which each rm is exposed,
and whether the rm uses nancial hedges in managing its currency risk.
In addition, we extend the return horizon as part of our rm-specic
estimations of foreign exchange exposure. To our knowledge, this is the
rst study to incorporate recent UK disclosure data into the estimation
and explanation of post-euro foreign exchange exposure.
3. Data Description, Sample Selection and Methodology
This study takes a rm-specic approach to examining the relation
between stock returns and changes in bilateral exchange rates, and the
use of nancial hedges to manage such currency risk. The study uses a
unique dataset of UK accounting disclosures, and departs from prior
UK Multinationals Financial Currency-Hedge Techniques 223
r 2008 Blackwell Publishing Ltd.
research in two important ways. First, we are able to identify the
principal currency to which each rm is exposed, rather than relying
on the broad exchange rate indices prevalent in previous research.
Second, we are able to identify whether or not each rm uses a particular
nancial currency hedge in managing its rm-specic exposure to
changing exchange rates. To the extent that rms do eectively reduce
their currency risk with nancial hedges, we expect to nd a negative
relationship between the use of such hedges and the estimated exposure
to the principal bilateral rate.
3
Specically, we test the following
hypothesis (in alternative form).
H
a
: There is a negative association between the level of foreign exchange
exposure and the use of nancial currency hedge techniques.
The UK accounting disclosures used in this study come from FRS 13,
entitled Derivatives and Other Financial Instruments: Disclosures
(Accounting Standards Board, 1998).
4
The general objective of this
reporting standard (eective for reporting periods ending after March
1999) is to require publicly listed rms to provide nancial statement
disclosures that allow users to assess each rms objectives, policies and
strategies for using nancial instruments. More specically, FRS 13
narrative and numeric disclosures should be useful in understanding
each rms currency risk prole, including an analysis of the net monetary
assets and liabilities with reference to the corresponding principal func-
tional currency, and the impact of nancial instruments on such risks.
5
In evaluating the usefulness of FRS 13 disclosures for nine UK banks
in 1999, Woods and Marginson (2004) caution that the narrative
disclosures tend to be generic in nature, and the numerical disclosures
are not always complete. Similarly, Marshall and Weetman (2007)
provide evidence which suggests that non-nancial rms FRS 13
disclosures lack transparency, and thus fall short of the standards
objectives. With regard to currency risk in particular, Woods and
Marginson nd that most banks do clarify their exposure to changing
exchange rates, although the form of such disclosure varies across banks.
In light of the descriptive nature of their content analysis, Woods and
Marginson call for future research on the usefulness of FRS 13
disclosures.
To estimate and explain foreign exchange exposure using the
rm-specic disclosures from FRS 13, we selected a sample of UK
multinationals listed in the June 2001 FTSE 250 that operate in the
non-nancial sector.
6
Data on principal currencies, hedge techniques,
224 Stephen D. Makar and Stephen P. Human
r 2008 Blackwell Publishing Ltd.
and foreign sales were hand gathered from annual reports. All remaining
data were machine gathered from the Datastream and Compustat Global
Vantage databases. For rms to be included in the nal sample,
comparable data must be available for monthly returns and principal
bilateral exchange rates, as well as for annual data on nancial and non-
nancial hedge techniques.
7
The resulting sample consists of 44 rms, for
the 19992002 sample period.
8
Table 1 provides a description of the sample rms FRS 13 disclosures
of principal currencies and currency hedge techniques. These narrative
UK Multinationals Financial Currency-Hedge Techniques 225
Table 1. Sample DescriptionFRS 13 Disclosures, by Year
1999 2000 2001 2002
Panel A: Firm-specic principal currency
US dollar 25 28 28 29
Euro 16 14 14 14
Australian dollar 2 1 1 1
Canadian dollar 1
Swiss franc 1 1
Total sample 44 44 44 44
Panel B: Firm-specic hedge technique
Firms using nancial hedges only (FHEDGE51)
Foreign denominated debt (FDD) 10 11 15 19
Foreign currency derivatives (FXD) 0 3 1 4
Both nancial hedge techniques used 2 2 5 1
Firms using non-nancial hedges only (NFHEDGE51)
Align foreign-denominated revenues and costs (ALIGN) 4 3 3 2
Diversify operations geographically (DIVERSIFY) 0 0 0 0
Other rms
Use both types (FHEDGE5NFHEDGE51)
FDD and ALIGN 5 5 5 6
FDD and DIVERSIFY 1 0 0 1
FDD, FXD and DIVERSIFY 0 1 0 0
FXD and ALIGN 2 2 0 0
FXD, ALIGN and DIVERSIFY 0 0 1 0
Use neither type (FHEDGE5NFHEDGE50) 20 17 14 11
Total sample 44 44 44 44
Panel A details the number of rms identifying a particular currency to which it is most exposed,
in accordance with FRS 13 disclosure requirements. These principal currencies are used in
rm-specic estimates of foreign exchange exposure for the 44 sample rms, as summarized in
Table 2.
Panel B details the number of rms identifying a particular hedge technique that is used to
manage its currency risk, as part of their FRS 13 disclosures. These currency techniques are used
in formal hypothesis tests of eective nancial hedging for the 44 sample rms, as described in
Table 3. Our denition of FHEDGE is the same as the one used by Hagelin and Pramborg
(2004). That is, FHEDGE is equal to 1 if the rm used foreign-denominated debt or if it used
foreign currency derivatives or if it used both, and is equal to 0 otherwise. NFHEDGE is equal
to 1 if the rm aligns foreign denominated revenues and costs (ALIGN) or if the rm diversies
operations geographically (DIVERSIFY), and is equal to 0 otherwise.
r 2008 Blackwell Publishing Ltd.
and numeric disclosures typically were provided in the rms Operating
and Financial Review section of its annual report. Panel A details the
principal currencies by year for our 44 sample rms, which we used in
rm-specic estimates of foreign exchange exposure. While many of the
rms identify the euro as the currency to which they are most exposed,
the majority of our sample is exposed primarily to changes in the US
dollar. This prevalence of US dollar exposure identied in our sample
rms FRS 13 disclosures is comparable to recent survey data for Dutch
rms in the 19941998 pre-euro period (DeJong et al., 2006). Other
studies investigating the exposure of European rms to the US dollar
include Glaum et al. (2000) and Nguyen et al. (2007). It is interesting to
note, however, that in our 19992002 post-euro period of study, most
UK sample rms are not principally exposed to the euro.
9
Moving to panel B of Table 1, the specic currency hedge techniques
used by our 44 sample rms are summarized by year. These data are used
to explain cross-sectional variations in the foreign exchange exposure
arising from the principal currencies detailed in panel A. As seen in panel
B, all but a few rms either use nancial hedges (FHEDGE51) or do
not use any currency risk hedges (FHEDGE5NFHEDGE50). For
example, in 1999, only four rms strictly use non-nancial hedges
(FHEDGE50, NFHEDGE51). The particular non-nancial hedge
that is most prevalent for the UK sample rms is the alignment of
foreign-denominated revenues with operating costs (e.g., Martin et al.,
1999; Capsta and Marshall, 2005). Although less prevalent, a few rms
do use geographic diversication of operations as a non-nancial
currency risk hedge. By diversifying to countries whose currencies do
not track each other closely, a rm may oset its exposure to changing
exchange rates (e.g., Pantzalis et al., 2001). It is also interesting to note
that across the years, UK sample rms have increased their use of
foreign-denominated debt to hedge currency risk, with a corresponding
decrease in the number of rms not using either hedge type.
As reviewed in Section 2, prior studies have recommended that the
Adler and Dumas (1984) approach to estimating foreign exchange
exposure be improved by incorporating rm-specic bilateral exchange
rates. Accordingly, principal currency data from FRS 13 are used in
time-series ordinary least-squares (OLS) estimates of equation (1).
Alternatively, IMF trade-weighted exchange rate indices are used, as a
point of comparison with prior research.
R
it
b
0
b
1
CER
t
b
2
R
mt
e
it
1
226 Stephen D. Makar and Stephen P. Human
r 2008 Blackwell Publishing Ltd.
where R
it
is the monthly return of rm i in period t (either the 1-month or
the 12-month return horizon), CER
t
is the percentage change in
exchange rate (either a rm-specic bilateral exchange rate per FRS 13
disclosures or the IMF trade-weighted British pound exchange rate
index) for period t, and R
mt
is the percentage change in FTSE All
market index for period t. Although prior research has noted the
advantages of using rm-specic bilateral exchange rates (e.g., Ihrig,
2001), we include the IMF trade-weighted index as a point of compar-
ison. Likewise, we include both the 1-month and the 12-month return
horizons in light of prior research, which has reported a delayed market
response to exchange rate changes (e.g., Grien and Stulz, 2001). In
estimating equation (1) over the 12-month return horizon, we employ the
NeweyWest (1987) correction method.
In testing the above-stated hypothesis, that there is a negative
association between the level of foreign exchange exposure and the use
of nancial currency hedge techniques, the absolute value of the
estimated coecient on the CER variable (b
1
) from equation (1) is
used as the dependent variable in equation (2). Cross-sectional variations
in these principal-currency exposure estimates are explained in relation to
the nancial hedge techniques that each rm identies as being used to
manage such risk.
We control for each rms size, percentage of foreign sales, and the use
of non-nancial hedges, in light of prior research (e.g., DeJong et al.,
2006). While larger rms are more likely to be multinationals (e.g.,
Bodnar and Wong, 2003), they may also benet from scale economies in
foreign exchange derivative use, geographic diversication and other
currency hedges (e.g., Pantzalis et al., 2001). Thus, the sign of the
estimated coecient on the size control variable is indeterminate. In
contrast, it is expected that the sign of the estimated coecient on the
percentage of foreign sales will be positive (e.g., Jorion, 1990), while the
sign on the non-nancial hedge indicator variable will be negative if such
hedging is eective in reducing the absolute level of currency risk (e.g.,
Pantzalis et al., 2001).
FXexposure
in
a b Independent Variable
in
e
in
2
where the dependent variable FX exposure is the absolute value of the FX
exposure estimate of rm i in period n (n 519992002) for the 1-month
return horizon. Cross-sectional OLS estimates of equation (2) include two
indicator variables (each independent variable equals 1 if the indicated
UK Multinationals Financial Currency-Hedge Techniques 227
r 2008 Blackwell Publishing Ltd.
condition is met in one or more rm years in period n) and two continuous
variables (each independent variable equals the rms time-series mean for
period n). In particular, the independent variable FHEDGE is an indicator
variable for the use of nancial hedges by rm i in period n (equals 1 if
annual report disclosures indicate use of either foreign exchange deriva-
tives or foreign-denominated debt to hedge currency risk, and 0 other-
wise).
10
The independent variable SIZE is the natural log of the market
value of common equity for rm i in period n. The independent variable
FSTS is the percentage of foreign sales for rm i in period n. Finally, the
independent variable NFHEDGE is an indicator variable for the use of a
non-nancial hedge by rm i in period n.
11
4. Results
Table 2 presents OLS regression results and univariate statistics for our
sample of 44 UK multinationals over 19992002. Panel A details the
foreign exchange exposure estimates corresponding to equation (1), while
panel B describes the general characteristics of rms using (or not using)
nancial hedges of currency risk. As seen in panel A, the use of rm-
specic bilateral exchange rates results in a greater number of signicant
exposure estimates (at the two-sided 10% level) compared with the IMF
exchange rate index, regardless of the time horizon. As expected, the
number of rms exhibiting signicant exposures increases with the return
horizon (e.g., Bodnar and Wong, 2003). In addition, the percentage of
signicant exposure estimates using equation (1) is more than double the
rate documented in prior research, for UK multinationals (e.g., Dom-
inguez and Tesar, 2006) or using rm-specic exchange rate indices (e.g.,
Fraser and Pantzalis, 2004). Using the FRS 13 disclosures to identify
each rms principal currency, we nd more UK multinationals with
signicant exposure.
Of the 25% of sample rms with signicant rm-specic exposure
estimates for the 1-month horizon, nearly two-thirds have negative
coecients on the equation (1) CER variable. This suggests that UK
multinationals rm value declines (increases) with the appreciation
(depreciation) of the UK pound against the principal currency. For
example, the majority of our sample rms identify the US dollar as their
principal currency (as detailed in Table 1), where the UK pound
appreciated against the US dollar over much of the sample period. In
addition, such multinational business activity makes up a substantial
portion of our sample rms sales. As detailed in panel B, the median
228 Stephen D. Makar and Stephen P. Human
r 2008 Blackwell Publishing Ltd.
UK Multinationals Financial Currency-Hedge Techniques 229
Table 2. OLS Regression and Univariate TestsFX Exposure Estimates and
Determinants
Panel A: FX exposure
estimates (OLS coecients,
signicant at the 10% level) # Firms % Firms # Positive # Negative
1-month return horizon (Eq. 1)
Firm-specic exchange rate 11 25.0% 4 7
IMF exchange rate index 3 6.8% 0 3
12-month return horizon (Eq. 1)
Firm-specic exchange rate 31 70.5% 16 15
IMF exchange rate index 5 11.4% 2 3
Panel B: FX exposure
determinants (millions of
British d) Mean
Standard
deviation Median
Wilcoxons
p-value
Firms using nancial hedge (Eq. 2)
SIZE 6.495 0.525 6.551 0.109
FSTS 0.513 0.300 0.521 0.636
Firms not using nancial hedge (Eq. 2)
SIZE 6.307 0.487 6.239
FSTS 0.450 0.260 0.487
Equation (1) for the 1-month return horizon: R
it
5b
0
1b
1
CER
t
1b
2
R
mt
1e
it
where R
it
is the
monthly return of rm i in period t (t 51 month), CER
t
is the percentage change in exchange
rate (either rm-specic bilateral exchange rate per FRS 13 disclosures or IMF trade weighted
exchange rate index) for period t, and R
mt
is the percentage change in FTSE All market index for
period t.
Equation (1) for the 12-month return horizon: R
it
5l
0
1l
1
CER
t
1l
2
R
mt
1e
0
it
where R
it
is the
annual return of rm i in period t (t 512 months), CER
t
is the annual percentage change in
exchange rate (either rm-specic bilateral exchange rate per FRS 13 disclosures or IMF trade
weighted exchange rate index) for period t, and R
mt
is the annual percentage change in FTSE All
market index for period t.
Equation (2): FX exposure
in
5a1b Independent Variable
in
1e
in
where the dependent variable
FX exposure is the absolute value of the FX exposure estimate of rm i in period n (slope
coecient from OLS estimates of equation 1 in n 519992002, for the 1-month return horizon),
the independent variable SIZE is the natural log of the market value of common equity for rm i
in period n, and the independent variable FSTS is the percentage of foreign sales for rm i in
period n.
Panel A details the number and sign of statistically signicant FX exposure estimates, at the
two-sided 10% level. FX exposure coecient estimates for the 44 sample rms in 19992002 use
OLS on Equation (1), and the NeweyWest (1987) correction for estimations using overlapping
observations. Panel B describes the FX exposure determinants that are used in tests of eective
nancial hedging (foreign exchange derivatives or foreign denominated debt use) for the 44
sample rms using OLS on Equation (2). Non-parametric tests of median dierences between
the two sub-sample FX exposure determinants are also provided.
OLS, ordinary least-squares.
r 2008 Blackwell Publishing Ltd.
percentage of foreign sales is 52.1% (48.7%) for rms using (not using)
nancial hedges. However, neither the median dierence in the percen-
tage of foreign sales nor size is statistically signicant, as indicated by the
Wilcoxon p-values.
Moving to the tests of the hypothesis, Table 3 summarizes the OLS
estimations of equation (2) using the 1-month return horizon with rm-
specic exchange rates.
12
The FHEDGE coecient is negative and
statistically signicant (at the one-sided 10% level or better), as expected.
These results indicate that UK multinationals eectively use nancial
currency hedges to reduce their exposure to changes in principal
230 Stephen D. Makar and Stephen P. Human
Table 3. OLS RegressionTests of Hypothesis 1-Month Return Horizon
with Firm-Specic ER: FX exposure
in
5a1b Independent Variable
in
1e
in
Model
Independent variables
Intercept FHEDGE SIZE FSTS NFHEDGE
Model 1 Estimate 0.404 0.272
p-value 0.006
(a)
0.135
(b)
Adjusted R
2
50.031
Model 2 Estimate 0.020 0.283 0.061
p-value 0.986 0.130
(b)
0.728
Adjusted R
2
50.010
Model 3 Estimate 0.002 0.282 0.067 0.044
p-value 0.999 0.138
(b)
0.715 0.896
Adjusted R
2
50.016
Model 4 Estimate 0.190 0.331 0.031 0.034 0.250
p-value 0.868 0.089
(a)
0.866 0.917 0.240
Adjusted R
2
50.004
Number of observations 544 rms.
As depicted in model 4, equation (2): FX exposure
in
5a1b Independent Variable
in
1e
in
where
the dependent variable FX exposure is the absolute value of the FX exposure estimate of rm i
in period n [slope coecient from OLS estimates of equation (1), n 519992002, using rm-
specic bilateral exchange rate], the independent variable FHEDGE is an indicator variable for
the use of nancial hedges by rm i in period n (equals 1 if annual report disclosures indicate use
of foreign exchange derivatives or foreign denominated debt to hedge currency risk, and 0
otherwise), the independent variable SIZE is the natural log of the market value of common
equity for rm i in period n, the independent variable FSTS is the percentage of foreign sales for
rm i in period n, and the independent variable NHEDGE is an indicator variable for the use of
a non-nancial hedge by rm i in period n (equals 1 if FRS 13 annual report disclosures indicate
the alignment of foreign-denominated revenues with operations costs or the geographic
diversication of operations to hedge currency risk, and 0 otherwise).
Table 3 summarizes tests of the hypothesis, that there is a signicant and negative relationship
between the level of FX exposure and the use of nancial hedges (foreign exchange derivatives
or foreign-denominated debt or both), using OLS on equation (2).
(a)
Signicant at the one-sided 5% level.
(b)
Signicant at the one-sided 10% level.
OLS, ordinary least-squares.
r 2008 Blackwell Publishing Ltd.
exchange rates. As indicated in models 14, the evidence of eective
hedging is robust to rm size, as well as to the percentage of foreign sales
and the use of non-nancial hedges.
13
Similar FHEDGE results are
obtained when the IMF trade-weighted exchange rate index is used in
place of the rm-specic bilateral exchange rate.
5. Conclusion
This study uses a unique dataset from recently available UK annual
report disclosures to take a rm-specic approach in examining the
relation between stock returns and changes in exchange rates. Unlike
prior studies, we are able to identify the principal currency to which each
rm is exposed, and whether or not the rm uses a particular nancial
currency hedge in managing its exposure to changing exchange rates. To
our knowledge, this is the rst study to incorporate such post-euro
accounting disclosure data into the estimation and explanation of foreign
exchange exposure.
The results presented in this study support the conclusion that UK
multinationals do make eective use of nancial currency-hedge techni-
ques to reduce the currency risk associated with changes in the bilateral
exchange rate to which they are most exposed. Incorporating such rm-
specic principal currencies into the estimation of foreign exchange
exposure, we document a greater number of rms with signicant
exposures, compared with the exposure estimates using a broad exchange
rate index. Consistent with prior research, we also nd that the number
of rms with signicant exposure to principal currencies increases with
the return horizon. Furthermore, the primary analysis indicates that our
hypothesis tests are robust to rm size, the percentage of foreign sales,
and the rms use of non-nancial hedges.
Although the results of our research are consistent with the hypothe-
sized negative association between exposure to changes in principal
currencies and the use of either foreign currency derivatives or foreign-
denominated debt, our ndings are limited by several factors. Our sample
is comprised of 44 large UK multinationals; thus, the results may not be
generalized to all UK rms or to rms outside the UK. Because of the
limited number of sample rms with data sucient to estimate and explain
foreign exchange exposure in the 19992002 period, we cannot directly
isolate the eectiveness of particular nancial or non-nancial hedge
techniques. For example, future research may distinguish the eectiveness
of foreign currency derivatives and foreign-denominated debt.
UK Multinationals Financial Currency-Hedge Techniques 231
r 2008 Blackwell Publishing Ltd.
Notes
1. Other researchers have used an alternative approach that examines the sensitivity of
cash ows or income and pricing to exchange rate changes (e.g., Walsh, 1994; Garner and
Shapiro, 1998; Bodnar et al., 2002, respectively).
2. For a comprehensive survey of the extant empirical evidence on the foreign exchange
exposure of non-nancial rms, see Bartram and Bodnar (2007).
3. Although we study the relationship between rm-specic estimates of foreign
exchange exposure and their use of nancial hedges, we do not examine the decision to
use foreign-denominated debt or foreign currency derivatives in the rst place. Froot et al.
(1993) provide a theoretical framework for such risk management practices.
4. FRS 13 was issued in September 1998, and has subsequently been withdrawn on
implementation of FRS 25, which was eective from January 2005 (Accounting Standards
Board, 2004). FRS 25 Financial Instruments: Disclosure and Presentation implements
the international standard IAS 32, in accordance with the September 2002 Accounting
Regulation that requires all EU listed rms to follow International Financial Reporting
Standards as of 2005.
5. FRS 13 requires rms to identify the principal functional currency of operations and
the associated net monetary assets and liabilities in order to allow users to understand the
exposures that result in adjustments reported in the prot and loss account (Accounting
Standards Board, 1998, paragraphs 3435). With regard to hedging this risk with nancial
instruments, FRS 13 requires rms to describe the transactions that have been hedged
(Accounting Standards Board, 1998, paragraph 21). Disclosures of the objectives and
strategies for issuing or holding such nancial instruments typically also should encom-
pass the risk management policies for using foreign-denominated debt and other hedging
techniques (Accounting Standards Board, 1998, paragraph 15).
6. Although Mallin et al. (2001) present survey evidence that nancial rms are major
users of foreign currency derivatives, we exclude nancial sector rms so that our analyses
are comparable with prior research.
7. Of the 187 rms in the June 2001 FTSE 250 that operate in the non-nancial sector,
55 rms lacked monthly returns, 37 rms did not operate outside the United Kingdom,
and 34 rms did not clearly identify their principal currencies or hedge techniques for all
sample years. Such evidence of incomplete disclosures is consistent with descriptive studies
of FRS 13 reporting practices (e.g., Woods and Marginson, 2004). Similarly, 16 rms
lacked complete disclosure of foreign sales. Finally, one rm was identied as an outlier
with fundamentally dierent economic characteristics, and thus was omitted from all
analyses. Our sample size of 44 UK rms is comparable in size to several other studies
using similar methodology. For example, DeJong et al. (2006) use a sample of 47 Dutch
rms, Glaum et al. (2000) have 71 German rms in their sample, He and Ng (1998)
investigate 45 Japanese rms, and Choi and Prasad (1995) examine 61 US rms.
8. With regard to industry membership, the 44 rms operate in the construction sector
(5 rms), the manufacturing sector (19 rms), the transportation sector (8 rms), the retail
sector (3 rms) and the service sector (9 rms).
9. For additional study of the impact of the euro on foreign exchange rate exposure, see
Bartram and Karolyi (2006). Using an event study approach centered on the euros
introduction, Bartram and Karolyi provide evidence that the change in exposure for rms
operating in 18 countries was statistically and economically small. Nguyen et al. (2007) use
methodology similar to Jorion (1990) for French rms, and provide evidence of reduced
exposure subsequent to the euros introduction.
10. Our denition of FHEDGE is the same as the one used by Hagelin and Pramborg
(2004). That is, FHEDGE is equal to 1 if the rm used foreign-denominated debt or if it
232 Stephen D. Makar and Stephen P. Human
r 2008 Blackwell Publishing Ltd.
used foreign currency derivatives or if it used both. We use FHEDGE in Table 1 and
include it as one of the independent variables in equation (2) when FHEDGE equals 1 in
one or more rm years during the 19992002 sample period.
11. NFHEDGE is equal to 1 if the rm aligns foreign-denominated revenues and costs
(ALIGN) or if the rm diversies operations geographically (DIVERSIFY), and is equal
to 0 otherwise.
12. Tests of the hypothesis employ the 1-month return horizon, given the extant
empirical evidence that nancial hedges eectively reduce short-term currency risk (e.g.,
Human and Makar, 2004). With regard to model assessment, diagnostics used to assess
OLS error-term assumptions included residual and normal probability plots, the Shapiro
Wilk test for non-normality, and the White test for heteroskedasticity. Diagnostics from
the INFLUENCE option in SAS were used to evaluate the presence of outliers. In total,
these model diagnostics suggest that departures from OLS assumptions do not hinder
hypothesis tests.
13. Beyond the sensitivity checks detailed in Table 3, we also veried that the primary
results are robust to increasing the return horizon from 1 to 12 months in dependent
variable estimates of FX exposure, or to including leverage as an explanatory variable
(total debt to total assets) or to using an alternative measure of the size control variable
(log of the book value of total assets).
References
Accounting Standards Board. Derivatives and Other Financial Instruments: Disclosures
(London: Accounting Standards Board, 1998).
Accounting Standards Board. Financial Instruments: Disclosure and Presentation
(London: Accounting Standards Board, 2004).
Adler, M. and B. Dumas, Exposure to Currency Risk: denition and Measurement,
Financial Management (1984), pp. 4150.
Allayannis, G. and E. Ofek, Exchange Rate Exposure, Hedging, and the Use of Foreign
Currency Derivatives, Journal of International Money and Finance (2001), pp. 273296.
Amihud, Y., Exchange Rates and the Valuation of Equity Shares, in R. Levich (ed.),
Exchange Rates and Corporate Performance (New York: Irwin, 1994), pp. 4959.
Bartov, E. and G. Bodnar, Firm Valuation, Earnings Expectations, and the Exchange-
Rate Exposure Eect, The Journal of Finance (1994), pp. 17551785.
Bartram, S. and G. Bodnar, The Exchange Rate Exposure Puzzle, Managerial Finance
(2007), pp. 642666.
Bartram, S. and G. Karolyi, The Impact of the Introduction of the Euro on Foreign
Exchange Rate Risk Exposure, Journal of Empirical Finance (2006), pp. 519549.
Bodnar, G., B. Dumas and R. Marston, Pass-Through and Exposure, The Journal of
Finance (2002), pp. 199231.
Bodnar, G. and M. Wong, Estimating Exchange Rate Exposure: issues in Model
Structure, Financial Management (2003), pp. 3567.
Capsta, J. and A. Marshall, International Cash Management and Hedging: a
Comparison of UK and French Companies, Managerial Finance (2005), pp. 1834.
Choi, J. and A. Prasad, Exchange Risk Sensitivity and its Determinants: a Firm and
Industry Analysis of US Multinationals, Financial Management (1995), pp. 7788.
DeJong, A., J. Ligterink and V. Macrae, A Firm-Specic Analysis of Exchange-Rate
Exposure of Dutch Firms, Journal of International Financial Management and
Accounting (2006), pp. 128.
Dewenter, K., R. Higgins and T. Simin, Can Event Study Methods Solve the Currency
Exposure Puzzle?, Pacic-Basin Finance Journal (2005), pp. 119144.
UK Multinationals Financial Currency-Hedge Techniques 233
r 2008 Blackwell Publishing Ltd.
Doidge, C., J. Grin and R. Williamson, Measuring the Economic Importance of
Exchange Rate Exposure, Journal of Empirical Finance (2006), pp. 550576.
Dominguez, K. and L. Tesar, Exchange Rate Exposure, Journal of International
Economics (2006), pp. 188218.
Donnelly, R. and E. Sheehy, The Share Price Reaction of UK Exporters to Exchange
Rate Movements: an Empirical Study, Journal of International Business Studies (1996),
pp. 157165.
Fa, R. and A. Marshall, International Evidence on the Determinants of Foreign
Exchange Rate Exposure of Multinational Corporations, Journal of International
Business Studies (2005), pp. 539558.
Fraser, S. and C. Pantzalis, Foreign Exchange Exposure of US Multinational Corpora-
tions: a Firm-Specic Approach, Journal of Multinational Financial Management
(2004), pp. 261281.
Froot, K., D. Scharfstein and J. Stein, Risk Management: coordinating Corporate
Investments and Financing Policies, Journal of Finance (1993), pp. 16291658.
Garner, C. and A. Shapiro, A Practical Method of Assessing Foreign Exchange Risk,
Midland Corporate Finance Journal (1998), pp. 617.
Glaum, M., M. Brunner and H. Himmel, The DAX and the Dollar: the Economic
Exchange Rate Exposure of German Corporations, Journal of International Business
Studies (2000), pp. 715724.
Grien, J. and R. Stulz, International Competition and Exchange Rate Shocks: a Cross-
Country Industry Analysis of Stock Returns, Review of Financial Studies (2001), pp.
215241.
Hagelin, N. and B. Pramborg, Hedging Foreign Exchange Exposure: risk Reduction
from Transaction and Translation Hedging, Journal of International Financial Man-
agement and Accounting (2004), pp. 120.
He, J. and L. Ng, The Foreign Exchange Exposure of Japanese Multinational Corpora-
tions, The Journal of Finance (1998), pp. 733753.
Human, S. and S. Makar, The Eectiveness of Currency-Hedging Techniques Over
Multiple Return Horizons for Foreign-Denominated Debt Issuers, Journal of Multi-
national Financial Management (2004), pp. 105115.
Ihrig, J.. Exchange-Rate Exposure of Multinationals: focusing on Exchange Rate
Issues. Board of Governors of the Federal Reserve System, International Finance
Discussion Paper No. 709, 2001.
Jorion, P., The Exchange-Rate Exposure of US Multinationals, Journal of Business
(1990), pp. 331345.
Kim, Y., I. Mathur and J. Nam, Is Operational Hedging a Substitute for
or a Complement to Financial Hedging?, Journal of Corporate Finance (2006), pp.
834853.
Mallin, C., K. Ow-Yong and M. Reynolds, Derivatives Usage in UK Non-Financial
Listed Companies, European Journal of Finance (2001), pp. 6391.
Marshall, A. and P. Weetman, Modelling Transparency in Disclosure: the Case of
Foreign Exchange Risk Management, Journal of Business Finance and Accounting
(2007), pp. 705739.
Martin, A., J. Madura and A. Akhigbe, Economic Exchange Rate Exposure of
US-Based MNCs Operating in Europe, The Financial Review (1999), pp. 2136.
Muller, A. and W. Verschoor, Asymmetric Foreign Exchange Risk Exposure: evidence
from US Multinational Firms, Journal of Empirical Finance (2006), pp. 495512.
Newey, W. and K. West, A Simple Positive Semi-Denite, Heteroscedasticity
and Autocorrelation Consistent Covariance Matrix, Econometrica (1987), pp.
703708.
234 Stephen D. Makar and Stephen P. Human
r 2008 Blackwell Publishing Ltd.
Nguyen, H., R. Fa and A. Marshall, Exchange Rate Exposure, Foreign Currency
Derivatives, and the Introduction of the Euro: French Evidence, International Review
of Economics and Finance (2007), pp. 563577.
Pantzalis, C., B. Simkins and P. Laux, Operational Hedges and the Foreign Exchange
Exposure of US Multinational Corporations, Journal of International Business Studies
(2001), pp. 793812.
Walsh, E., Operating Income, Exchange Rate Changes, and the Value of the Firm: an
Empirical Analysis, Journal of Accounting, Auditing and Finance (1994), pp. 703724.
Williamson, R., Exchange Rate Exposure and Competition: evidence from the Auto-
motive Industry, Journal of Financial Economics (2001), pp. 441475.
Woods, M. and D. Marginson, Accounting for Derivatives: an Evaluation of Reporting
Practices by UK Banks, European Accounting Review (2004), pp. 373390.
UK Multinationals Financial Currency-Hedge Techniques 235
r 2008 Blackwell Publishing Ltd.

You might also like