Docstoc

A Market Based Study of Foreign Exchange Exposure – The Effect of

Document Sample
A Market Based Study of Foreign Exchange Exposure – The Effect of Powered By Docstoc
					        A Market Based Study of Foreign Exchange Exposure –

            The Effect of Time Horizon and Survivorship Bias




                                     Tony Muff*
                              University of Northampton

                                  Stephen Diacon
                                  Margaret Woods
                              University of Nottingham




* Correspondence Details

Dr Tony Muff
Northampton Business School
University of Northampton
Park Campus
Northampton NN2 7AL
United Kingdom

T 01604 892356; F 01604721214; tony.muff@northampton.ac.uk

                                          1
                A Market Based Study of Foreign Exchange Exposure –

                   The Effect of Time Horizon and Survivorship Bias


Abstract

The overall impact of exchange rates on firms is extremely complex especially as indirect
exposures are difficult to identify and measure. Consequently a large number of market
based studies have failed to identify any residual impact of exchange rate volatility on the
firm’s market value. There are various possible explanations for these findings, including
the time horizons selected for the research and the sample selection process, both of
which are addressed in this paper.

We examine the existence of a contemporaneous relationship between the stock returns of
UK non-financial firms and fluctuations in foreign exchange rates using three sub periods.
The sample was split into three equal periods covering the years 1985-1989, 1990-1994,
and 1995-1999. The findings show the market values of approximately 15% of firms are
exposed at the 10% level to exchange rate volatility. There are also a high proportion of
positive exposures among the sample firms with significant exchange rate exposures,
indicating a higher proportion of firms benefit from an appreciation of the £. The findings
of this subset analysis indicate that exposure to exchange rate movements are not stable
over time and increase in periods of economic uncertainty.


Secondly, the study examines the impact of survivorship bias that can arise as firms drop
out of the sample over the three periods. In order to test and correct for survivorship bias
a probit sample selection model is used. The results indicated that survivorship bias is a
problem in times of economic uncertainty: survivorship bias was significant during the
period 1990-1994 where the UK economy exchange rates and interest rates were volatile.


We conclude that the extent and significance of foreign exchange rate exposure is low and
consistent with the reported findings of earlier studies. However, the full extent of firm
exposure to exchange rate shocks may be better understood if researchers consider
survivorship bias in their sample data sets.




                                             2
1. Introduction


It is widely recognized that changes in exchange rates influence firm value through an
impact on unexpected or unanticipated cash flows. These effects are traditionally classified
into transaction or direct exposure (associated primarily with short-term risk) and
economic or indirect exposure which is associated with longer term fluctuations - which
are less easy to specify and hedge (Chow, Lee and Solt, 1997). However the lack of strong
empirical association between exchange fluctuations and stock returns appears perplexing
(Nance, Smith and Smithson, 1993) even when the construction of the sample has
concentrated only on major exporters (Bartov and Bodnar, 1994; Donnelly and Sheehy.
1996)1.

The failure of many studies to establish significant exposure of firm value to exchange rate
volatility may be as a result of the varying treatment of a number of important factors as
the time horizons, the portfolio construction, and the sample selection process (see
Bartram and Bodnar 2007 for a review). These contrasting positions have contributed
largely to ambiguous evidence on exchange rate exposure. Furthermore doubt has also
been on the use of a composite (trade weighted) index to measure exposure because such
an index is unlikely to capture the exchange rate exposure of any individual firm (Miller
and Reuer, 1998; Martin, Madura and Akhigbe, 1999; Ihrig and Proir, 2004; Bartov and
Bodnar, 2004; and Parsley and Popper, 2006). Recently, Salifu, et al. (2007) examined
Ghanian firms exposure to the dollar and the £ bilateral exchange rates and report 55% of
sample firms exposed to the $ and 35% to the £.

This study pays particular attention to two reported areas of areas of ambiguity, the time
horizon of the studies and possible bias arising from the sample selection process.

As the variability of the exchange rate exposure can have significant short term impact on
firm value (Jorion 1990; Glaum et al. 1999) a number of researchers including Choi and
Prasad (1995), Griffin and Stulz (2001), Shin and Soenin (1999) and Dominguez and
Tesar, (2004) have examined shorter sub periods that capture exchange rate shocks
caused by changes in macro economic circumstances and to determine if these vary over
macro economic cycles. The general conclusions from these studies point to levels of
instability in results over time and taking a shorter time horizon would offer greater clarity
to a firm’s currency exposure. In most studies that consider sub period analysis, the
exposure coefficients and their significance vary noticeably across sub periods (Glaum et
al., 1990). For example, Choi and Prasad (1995) employed two sub periods representing
a strong and weak US$ period respectively. The effects vary during the period but
evidence suggests that the number of US firms with significant exchange rate exposures
was higher during the weak Dollar period. Griffin and Stulz (2001) studied high and low
$/Yen volatility, and found that exposure for a Japanese firm was more pronounced during
the period when the Yen appreciated sharply. In the UK, Donnelly and Sheehy (1995)
formulated sub periods around the UK entry to the ERM and found that exposure of UK
firms fell during membership at the time when the £/€ exchange rate volatility fell. El
Masry (2007) analyzed UK firms at industry level and reported that the significance of
exchange rate exposure coefficients declined during the period when the UK was a
member of the ERM.




1
  Recent investigations into the empirical relationship between stock returns and exchange rate movements in the
US have confirmed absence of significant association (Jorion 1990; Bodnar and Gentry 1993; Amihud 1994; Bartov
and Bodnar 1994; Khoo 1994; Choi and Prasad 1995; Prasad and Rajan 1995; Gao, (2000); Griffin and Stulz
(2001).

                                                       3
Dominguez and Tesar (2004) measure the relationship between exchange rate volatility
and firm value in eight industrialized and developing countries over a 20 year time span
and a broad sample of firms. They conclude that exposure is not stable and hypothesize
that this is due to the adaptability of firms to take advantage of exchange rate volatility by
actively managing imported input share and export share and using adaptive pricing
strategies and the value of mark-up.

The problems with survivorship bias arise because of the sample selection process as the
analysis can only include firms which continue to trade over the sample period. Studies
conducted at the level of the individual firm use data for firms which are live throughout
the sample period, and exclude non-surviving firms. This could potential cause a serious
selection bias if firms are included in a nonrandom way. Kothari, Shanken, and Sloan
(1995) and Brown and Goetzmann, (1995) cite survival bias as a problem that can
exaggerate predictive power2. In the case of exchange rate exposure, survivorship bias
can arise if the firms which drop out of the sample do so because of their inability to
manage exchange rate risk. Exchange rate exposure studies typically use long-term time
series to determine currency exposure and only include firms that are live throughout the
duration of the period. Potential survivorship bias arises because these surviving firms
may have developed a level of sophistication in their exchange rate risk management
sufficient to avoid significant currency exposures and potential financial; distress,
liquidation or takeover.

Another potential selection bias exists in cross sectional explanatory variables. Davis
(1996) identifies a problem of spurious explanatory power of variables. For example, if
firms with high book to market ratios are typically distressed, then firms that turn
themselves around may experience high stock returns, while those that remain distressed
may fail. If the former are included in a sample and the latter excluded, then firms with
high book to market ratios may have high subsequent returns, precisely because the non-
surviving firms are excluded. In effect, accounting variables that identify exposure due to
financial distress may be under reporting the extent of exposure, as they are
representatives of firms that have effectively controlled for exposure.


The rest of the paper proceeds as follows. Section 2 provides a brief overview of the
determinants of foreign exchange exposure and identifies those variables that might
explain inter-firm variation. Section 3 discusses the methodological issues, and focuses
particularly on the treatment of the firm’s choice of exchange rate exposure before going
on to examine the twin issues of time horizon and selection. Section 4 reviews the
determinants of cross-sectional variation in exchange rate betas over three different time
horizons, which reflect substantially different levels of exchange rate volatility associated
with the UK’s membership of (and then departure from) the European Exchange Rate
Mechanism. Section 5 then explores the effect of possible selection bias by identifying
differences between firms which survive between two consecutive time periods and those
which do not. Section 6 provides a short conclusion.


2. The Determinants of Foreign Exchange Exposure

Donnelly and Sheehy (1996) argue that a firm’s exposure to exchange rate risk would be
enhanced by greater openness in the economy. Several studies have supported this
hypothesis including Bodnar and Gentry (1993), He and Ng (1998), De Jong et al. (2002)
and Kiymaz (2003). Nydahl (1999) studying the exchange rate exposure of Swedish firms
with foreign sales ratio of at least 10% found 19 out of 47 firms (40%) are negatively

2
 For further information on selection bias in UK databases see Nagal, S., (2001), Accounting information free of
selection bias. A new UK database. London Business School Working Paper
                                                       4
exposed to exchange rate risk at the 10% level.

Bartov and Bodnar (1994) and Chow et al (1997) argue that effective foreign currency
hedging will tend to introduce a downward bias in the estimated exposure coefficients and
reduce the power of the results. This is because the models used only measure residual
exposure after hedging policies have taken place, therefore it is not possible to fully
determine the exposure of firms to exchange rate shocks, and only ex post hedging activity
data is available. Firms themselves can employ natural hedging techniques by matching
assets and liabilities in the same currency, and by diversifying operations - so reducing
exposure to insignificant levels. The reported hedging activities used by researchers are
largely financial hedging instruments such as forwards, options and swaps and it is difficult to
obtain any reported data on the geographical distribution of both revenues, and costs and
this reduces the power of the results (Geczy, et al., 1997; Allayannis and Ofek, 2001,
Tufano, 1996).

The most obvious and potentially important contingent variable is the firm’s involvement
in international economic activity. To measure the level of international operations,
previous studies have typically used involvement in exporting (Jorion, 1990; Donnelly and
Sheehy, 1996; He and Ng, 1998; Bradley and Moles, 2002). However, a common problem
in these surveys is the difficulty in observing import exposure, so that the net extent of
exposure is therefore unobservable. Jorion (1990), Choi and Prassad (1995) and
Allayannis and Ofek (2001) report the level of foreign sales and size are a key determinant
of exchange rate exposure of large US multinationals. The scale economies in hedging
have also created some debate: if large firms have more foreign activity relative to small
firms, they may have more exposure (He and Ng 1998, Bodnar and Wong 2000, Diodge et
al, 2002). However Chow, Lee, and Solt (1997) and Dominguez and Tesar (2001) support
the view that there are economies of scale in the employment of hedge management
products. The size variable therefore is a measure of the extent to which economies of
scale reflect the use of hedge management products.

Optimal hedging theories postulate that a firm’s hedging activities will reduce exposure to
currency fluctuations. Smith and Stulz (1985) and Nance et al. (1993) demonstrate that
exchange rate exposure is sensitive to financial distress and agency costs. Firms with
greater financial leverage are more likely to hedge and hence would be less exposed to
exchange rate risk. Agency cost and other asymmetric information and capital market
imperfections create a cost differential between internally generated funds and external
costs of finance. Froot et al. (1993) demonstrate that if there is a cost differential between
external sources of finance and internal generated funds, there is a rationale for hedging
to increase the certainty of cash flow. Froot et al. (1993) predict that there is an incentive
to increase liquid assets if firms are faced with a cost differential between internal and
external financing. The underinvestment cost hypothesis also suggests an interaction
between growth opportunities and costly external finance. The higher the growth
opportunities the greater the incentive to hedge if there is a dependency on costly
external finance.

In the presence of a convex tax schedule, firms would reduce expected taxes by stabilising
cash flows through hedging, therefore the greater the level of tax convexity the greater
the incentive to hedge (Nance et al. 1993). There has been some conjecture as to the
reliability and usefulness of the tax convexity argument and hedging (Graham and Rogers,
2002) and only a few researchers have found a significant association between the
motivation to use derivatives and the tax convexity hypothesis (Berkman and Bradbury,
1996, and Dunne et al. 2004). Dunne et al. (2004) on a study of 209 U.K. firms, report a
negative and significant relationship between the tax variable and the extent of disclosure
reporting. They suggest that firms that pay the lowest rate of tax on their profits have the
highest disclosures about derivatives usage because they are using these financial
instruments to minimise their tax liability.
                                               5
Low liquidity ratios imply higher bankruptcy risk (Froot et al. 1993) and higher agency
costs (Berkman and Bradbury, 1996; Howton and Perfect, 1998). Nance, et al . (1993)
report that by investing in hedging substitutes, namely liquid or less risky assets, the
probability of default can be reduced. Most prior studies, including He and Ng (1998) use
the quick asset ratio to proxy for these hedging substitutes.

Size and economies of scale in the use of derivatives, which involves significant fixed
costs, may lead to further cross sectional variation in exposure would be inversely related
to size. If there are economies of scale in hedging risk, then the level of exposure would
be inversely related to size, however there exists some ambiguity in the association
between size and exposure and therefore the relationship is subject to empirical
investigation.


3. Research Methodology

The basic model

In order to investigate the potential for any instability in the sensitivity of share returns
and exchange rate movements, the data set is sub divided into three periods, identified as
follows:

    •    the 5 years before UK entry into the ERM (January 1985-December 1989)
    •    a five year period during entry and the immediate aftermath after leaving the ERM
         (January 1990-December 1994)
    •    a further five year period post entry (January 1995-December 1999).

These periods are also reflective of structural breaks in exchange rate volatility, as
illustrated in Tables 3a and 3b.

To be included in this sample a firm had to be a constituent of the UK Financial Times All
Share Index between 01 January and 31 December for each sub period. To be included in
the sub period sample, each firm was required to report continuous data for the whole sub
period. Financial firms were excluded, yielding the total number of included firms in the
three sub-periods as 397, 446 and 419 respectively.

The typical method used to identify firm specific factors that identify exposure is firstly to
determine the degree of exposure using time series regression, then use the resulting
exchange rate beta as the dependent variable in a further cross section regression to
identify the firm specific determinants (for examples, see Miller and Reuer, 1998; He and
Ng, 1998;). However, managers can choose the extent to which they avoid exchange rate
risk by employing off balance sheet financial hedging instruments or by structuring their
assets and liabilities to reduce exposure (Bartov and Bodnar, 1994; Amihud, 1994).
Therefore, an aspect shared by many prior studies that have used the large sample survey
method is the problem of endogeneity. Although disclosure requirements have been
significantly extended in the UK through the provisions of Financial Reporting Standard 13
(FRS13), some exchange rate risk management practices remain unobservable.3 In effect,
the independent variables in the model are potentially choice variables correlated with

3
  FRS13, ‘Derivatives and other Financial Instruments: Disclosures’, published by the Accounting Standards
Board on September 24th 1998 requires all companies with a capital instrument listed on a stock exchange or
market in the UK to explain in financial statements the role that financial instruments play in their funding. The
provisions of FRS13 will allow this research to be extended in more detail in the future. At the time this research
was conducted there were insufficient annual reports available under FRS13 provisions to construct meaningful
proxies for empirical testing of numerical data.

                                                        6
unobservables. To deal with this bias it is appropriate to ask two questions in cross
sectional analysis. First, what are the determinants of the decision to become exposed?
Secondly, given the first decision, what factors moderate the level of risky activities in
those companies that have chosen to engage in them? It is likely that the situational and
firm specific variables discussed above will influence both aspects.


To determine the extent of firm currency exposure the regression model below was used
and applied to each sub period. The monthly returns index of share price was regressed
on the monthly all share returns index and the monthly change in four currency models.
The UK trade weighted exchange rate, the US Dollar, Japanese Yen, and Euro. In total,
there were 60 monthly observations for each period:

                     Ri,t = αi +βc0i CUR t + βRi Rmt +   εit                           [1]

where R it=the realised return for security i for period t, α = intercept; CURt = the
contemporaneous exchange rate change in period t; βc0i = the sensitivity of firm i‘s stock
returns to exchange rate changes in period t; R mt = the total market return for period t;
and βRi = the sensitivity of firm i‘s stock returns to changes in the market index.

A second stage cross sectional analysis is usually performed using βc0i as the dependent
variable in order to identify the significance of firm specific factor determinants of
exposure (for example, Jorion, 1990; He and Ng 1998). In order to explain inter-firm
differences in βc0i accurately, we need information on each firm’s exchange rate risk
management strategies – but this is unobservable. In the context of exchange rate
exposure and its determinants there are unobservable variables that influence the decision
to manage risk. Individual firms can choose to be exposed or not and the problem creates
the missing variable problem which can lead to bias. We choose to control this problem by
using a Heckman two-stage selection model where we assume that the firm selects
whether or not to be exposed to exchange rate risk by utilizing unobservable risk
management activity - which is reflected in a significant coefficient for the parameter βc0i
in [1]). The extent of the subsequent exposure (which could be positive or negative) is
then modeled as an outcome process, conditional on significant exposure in the selection
stage.


The Heckman selection model (Heckman, 1979) is based on two latent variable models.
The outcome regression model is given by:

                            |βc0i|* = xj β ′ + u1j                                      [2]

where |βc0i|* is defined as the absolute value of the exchange rate beta for firm j
conditional on it being significant at the 10% level (that is, |t| ≥ 1.65). xj is represents
the vector of regressors that determine exposure (as described in Table 1) and u1j the
error term. Following Choi and Prasad (1995), the dependent variable in the outcome
model [2] is taken as the absolute value of the exchange beta coefficient conditional on
10% significance |βc0i|*. The sign of the estimates βc0 is irrelevant as it merely reflects
whether the firm is a net exporter (βc0 < 0) or importer (βc0 > 0).


Secondly, the selection model identifies a firm’s decision to be exposed:

                            zjγ + u2j > 0                                               [3]

where u 1 ~ N(0,σ); u2 ~ N(0, 1); corr(u 1, u2) = ρ; and zj are regressors (which are
described in Table 1).

                                             7
When ρ = 0, ordinary least squares estimates of [2] provides unbiased estimates, but
when ρ ~= 0 these estimates are biased. More formally as long as the error term in the
selection equation is correlated with that in the outcome equation (because the attitude to
risk management is an unmeasured variable in each equation), the error term in the
selection model is not of mean zero and is correlated with explanatory variables. The
Heckman model then turns the selection bias problem into an omitted variables problem
and proposes a method for estimating the omitted variable and inserting it into the
outcome equation. What is of interest is the effect of unmeasured characteristics of the
respondents on the exposure decision not available in the coefficients of the explanatory
variables. However, this information is available in the residuals. In the Heckman
procedure, the residuals of the selection equation are used to construct a selection bias
control factor, given by the Inverse Mills Ratio λ:
                                           φ (z j γˆ )
                                    λj =                                                        [4]
                                           Φ (z j γˆ )
where   φ is a standard normal distribution and Φ is the cumulative standard normal
distribution.

Variables and Data Descriptives

The first step is to estimate a selection model [3] where zj are the selection variables
defined as book-to-market rank, capital gearing ratio, cash earnings per share, % tax
payable, quick asset ratio, export ratio, and size (as defined in Table 1). The expected
signs of explanatory variables in the selection model are given in Table 2.


   Table 1: Descriptive Statistics for Outcome and Selection Regression Models

Descriptive Statistics                                   Mean      Std. Dev. Min       Max
BMCAT= Average ratio of the net       Period 1           6.1460    2.6626    1         12
book value of assets to market        Period 2           6.2719    2.9439    1         12
capitalization. Ranked by %
                                      Period 3           5.6538    3.1691    1         12
CGR= Average ratio of book value of Period 1             27.0114   32.4114   0         593.72
long term borrowings to the year    Period 2             32.4230   51.2270   0         994.36
end market value of the equity
                                    Period 3             41.9864   164.0485 0          3324.38
CASHEPS= Average ratio of cash        Period 1           24.3010   79.1337   -59.73    1504.46
earnings per share                    Period 2           22.9731   32.0945   -264.29   399.82
                                      Period     3       30.8297   27.9452   -35.86    233.19
TAX= Total tax charge as a            Period     1       32.2758   27.7485   -123.94   403.66
percentage of pretax profit           Period     2       31.4062   48.5841   -218.52   601.22
                                      Period     3       27.0372   27.9627   -209.00   317.58
QAR= Quick asset ratio                Period     1       0.9800    0.4891    0.04      3.42
                                      Period     2       1.0163    0.5667    0.06      5.93
                                      Period     3       1.0121    0.8309    0.11      11.29
EXPORTS= Average percentage of        Period     1       29.2904   27.3109   0.00      97.25
exports to total sales of firm        Period     2       32.9493   29.1267   0.00      94.76
                                      Period     3       37.0455   30.4559   0.00      99.84
LNMV= Natural log of market value     Period     1       11.5003   1.6150    7.61      16.52
                                      Period     2       12.0159   1.6945    7.36      16.90
                                      Period     3       12.6517   1.7104    3.81      17.87

                                                  8
It would be expected that the book to market rank, a proxy for growth opportunities
(BMCAT) would be positively related to currency exposure. The higher the BMCAT the
lesser a firm’s incentive to employ currency derivatives to hedge in order to reduce
underinvestment cost. For a similar reasoning, we expect to find a negative relationship
between the gearing ratio (CGR) and exposure. Higher levels of financial distress would
encourage hedging. It is expected that firms with higher cash earnings per share
(CASHEPS) will be less sensitive to financial distress and agency cost of debt and therefore
a positive relationship between exposure and cash earnings is anticipated. The higher the
TAX variable the higher the total fixed cost to the firm and the lower the interest cover
ratio. Firms would wish to reduce the variability in tax and stabilise income through
hedging, and for this reason, a negative relationship between exposure and TAX is
predicted. Firms can use hedging substitutes as a buffer to carry any exchange rate
losses, especially if the exposure is difficult to quantify; therefore, a negative liquidity
value QAR would indicate that firms had insufficient liquidity to control any residual
exposures after formal hedging techniques had been exhausted. Firms with international
operations will have more reason to be concerned about exchange risk as evidence
suggests that exposure will rise with the level of international operations (EXPORTS).
Larger firms may be more able to protect risky assets by the extent of their international
operations and more informal hedging activities and therefore larger firms (LNMV) may
accept a higher level of exposure than their smaller counterparts.


                   Table 2 Firm-Specific Variables, Expected Sign

                        Variable     Selection    Outcome
                                     Model        Model
                      BMCAT          +ve          -ve
                      CGR            -ve          +ve
                      CASHEPS        +ve
                      NOL            -ve          +ve
                      TAX            -ve          +ve
                      QAR            -ve          +ve
                      EXPORTS        +ve          +ve
                      LNMV           +ve




The coefficients in the outcome model above should be interpreted as measuring the extent
to which firm-specific variables affect the impact on stock returns of the firm’s ‘residual’
exposure to exchange rate risk given that there is a significant non-zero exposure. Included
are variables deemed important in identifying exposure after all formal and informal
methods to reduce exposure have been exhausted; that is after all risk management
activity has taken place. The key determinants of exposure are financial distress,
underinvestment, taxation, and hedging substitutes. The variables used to proxy for these
determinants are book to market rank (BMCAT), capital gearing ratio (CGR), taxation (TAX)
and hedging substitutes (QAR) and to control for international operations (EXPORTS). All
other variables are excluded.




                                             9
4 The Effect of Different Time Horizons

Descriptive Statistics

Table 3 below presents descriptive information on the four exchange rate betas generated
from Equation [1]. Table 3(a) presents the raw data and Table 3(b) presents the absolute
value, and the negative and positive beta values. Conclusions from this descriptive data
suggest significant structural changes in currency values. The Euro was especially volatile
during the period 1990-1994 when the UK was a member of the exchange rate
mechanism.

The absolute values in βc0 reported in Table 3b give an indication of overall magnitude of
exposure to exchange rate movements. The most interesting finding is the increase in
volatility of the £/Euro during the mid period 1990-1994, a time when the UK was a
member of the Exchange Rate Mechanism. Prior evidence (Donnelly and Sheehy, 1995; El
Masry, 2007) indicates that exposure fell during British membership of the ERM. However
as this study includes a period to represent the transitional period after leaving the ERM,
the whole period has proved to be highly volatile. The main cause of the volatility is
reflected in the positive beta coefficients indicating that during this period most firms lost
when the Pound deteriorated against the €. This is very interesting and indicates that UK
firms are more vulnerable to costs rather than revenues.


 Table 3. Descriptive statistics for the dependent and explanatory variables: for
               three sub periods 1985-89, 1990-94 and 1995-99

Table 3(a) Beta value for the Exchange Rate Index (TWI), the Dollar, Yen and the Euro.
Raw data. In Period 1985-1989 there were 397 observations; 1990-1994, 445
observations and in 1995-1999, 416 observations


                         Period        Mean        Std Dev   Min         Max
           Raw Data
           TWI βc0        1985-1989    -.029187    .535380    -2.79210   1.63560
                          1990-1994     .262089    .766598    -2.51000   5.34000
                          1995-1999    -.081905    .741667    -2.87767   2.38353
           Dollar βc0     1985-1989     .003904    .343947    -1.45654   1.20771
                          1990-1994     .008157    .411018    -2.18000   2.19000
                          1995-1999    -.116845    .721548    -3.99552   3.58489
           Yen βc0        1985-1989    -.113808    .468400    -3.08911   1.92498
                          1990-1994     .036157    .420111    -1.74000   1.78000
                          1995-1999     .137364    .335002    -1.06132   1.17041
              Euro βc0    1985-1989    -.003526    .521682    -2.56772   2.01820
                          1990-1994     .363168    .826080    -1.97000   5.81000
                          1995-1999    -.139440     693289    -2.54462   2.07070




                                              10
 Table 3(b) Absolute, Negative and Positive Beta values for the Exchange Rate Index
 (TWI) the Dollar, Yen, and the Euro. In Period 1985-1989 there were 397 observations;
 1990-1994, 445 observations and in 1995-1999, 416 observations

                 Period       Absolute Beta         Negative Beta      Positive Beta
                              Mean     Std Dev      Mean       Std     Mean       Std Dev
TWI βc0          1985-1989     .385419   .372237     -0.401460 Dev
                                                               0.42739   0.368291 0.302650
                 1990-1994     .560803   .586373                     9
                                                     -0.415723 0.43657   0.638916 0.641137
                 1995-1999     .579373   .469374                     6
                                                     -0.611326 0.50392   0.541745 0.423341
Dollar βc0       1985-1989     .260351   .224412                     1
                                                     -0.276650 0.25967   0.246266 0.188237
                 1990-1994     .279513   .300449                     5
                                                     -0.293880 0.31750   0.268493 0.285136
                 1995-1999     .547377   .483707                     7
                                                     -0.600690 0.52985   0.481456 0.411626
Yen βc0          1985-1989     .336262   .345002                     9
                                                     -0.375374 0.38486   0.277716 0.265304
                 1990-1994     .300649   .294772                     3
                                                     -0.302461 0.29716   0.300280 0.293381
                 1995-1999     .280651   .228441                     7
                                                     -0.211375 0.18398   0.316160 0.240804
Euro βc0         1985-1989     .379397   .357574                     9
                                                     -0.400053 0.39625   0.360436 0.317765
                 1990-1994     .621135   .658262                     0
                                                     -0.402482 0.36687   0.718289 0.733792
                 1995-1999     .547118   .447303                     5
                                                     -0.615549 0.47756   0.460853 0.390417
                                                                     2


 Impact of Currency Fluctuations on Firm Exposure

 Table 4     a-c shows cross sectional distribution of UK firms βc0 for the three sub sample
 periods.    The Table reports the total number of negative and positive exposures and the
 number      of significant exposures obtained. The summary information indicates that on
 average     15% of sample firms are exposed to exchange rate risk at the 10% level.
 The overall level of exposure was most significant during sub period 2, i.e.1990-1994
 when the UK was a member of the ERM. The total number of significant positive exposures
 during this period exceeded negative exposures by over 300%. This suggests that during
 this period an appreciation/depreciation of £ had a significant positive/negative impact on
 stock returns and indicating that firms were exposed through costs rather than revenues.
 It would be expected therefore that exporters would benefit from a fall in the £ and
 importers of goods would face a significant increase in the costs of production that would
 have to be absorbed. The overwhelming significant positive exposure supports the
 argument that firm’s cash flows are more sensitive to cost drivers rather than revenue
 driven.

 During the 15-year period of study, there have been significant structural differences
 between the three sub periods and this has significantly affected the level of exposure of
 UK companies. This is most notable during period two whereas the overwhelming level of
 exposure was positive although the overall aggregate exposures suggest there is
 instability in the direction of exchange rate exposure, and the level and extent of this
 instability differs between bilateral exchange rates.




                                               11
    Table 4 Exchange Rate Risk Exposures of βc0, of the 397 Largest UK Firms
Table 4a Exchange Rate Risk Exposures of βc0, of the 397 Largest UK Firms 1985
– 89: Exchange rates are the BofE trade weighted index TWI, the Dollar, Yen and
Euro

          Significant   Significant   Significant   Significant   Negative   Positive
          Exposure      Exposure      Negative      Positive      Exposure   Exposure
          (5%           (10%          Exposure      Exposure
          level)        level)        (10%          (10%
                                      level)        level)

 TWI      35   (8.8)    49   (12.3)       22            27          205         192
 US$/£    36   (9.0)    61   (15.6)       25            36          184         213
 Yen/£    37   (9.3)    58   (15.1)       41            17          238         159
 Euro/£   30   (8.3)    52   (13.1)       27            25          190         207



Table 4b Exchange Rate Risk Exposures of βc0, of the 445 Largest UK Firms 1990
– 94: Exchange rates are the BofE trade weighted index, the Dollar, Yen and Euro

          Significant   Significant   Significant   Significant   Negative   Positive
          Exposure      Exposure      Negative      Positive      Exposure   Exposure
          (5%           (10%          Exposure      Exposure
          level)        level)        (10%          (10%
                                      level)        level)


 TWI      52   (11.6)   85   (19.1)       12            73          151         286
 US$/£    44   (9.8)    66   (15.0)       26            40          206         239
 Yen/£    38   (8.4)    57   (12.8)       21            36          195         250
 Euro/£   51   (11.6)   73   (16.4)        9            64          141         304




Table 4c. Exchange Rate Risk Exposures of βc0, of the 416 Largest UK Firms 1995
– 99: Exchange rates are the BofE trade weighted index, the Dollar, Yen and Euro

          Significant   Significant   Significant   Significant   Negative   Positive
          Exposure      Exposure      Negative      Positive      Exposure   Exposure
          (5% level)    (10%          Exposure      Exposure
                        level)        (10%          (10%
                                      level)        level)

 TWI      27   (6.5%)   52   13%)         32            20          225         191
 US$/£    25   (6%)     37   (9%)         28             9          230         186
 Yen/£    45   (11%)    67   (16.%)       10            57          141         275
 Euro/£   38   (9%)     56   (13%)        41            15          232         184




To measure instability in exposure, Table 5 summarizes the number of firms identified as
being exposed across more than one period and indicates where the sign of exposure
changes between periods. The results are quite marked for the Euro with 10 out of 19

                                               12
firms reporting a change in sign between two sub periods indicating a high level of
instability. Equally, there were no firms exposed across the whole period suggesting firms
are managing their exposures in the medium to long term. There results suggest that
structural difference in exposure is likely to have a significant influence on the short term
hedging strategy and exposure management of each firm especially if they are exposed to
multiple exchange rates. Although detailed information is not available through market
based studies, this study suggests firms are dynamically adjusting their exchange rate risk
management strategy to control for this exposure. 4


Although the exposures are higher than comparable US studies (Jorion 1990; Choi and
Prassad 1995), they do not offer a convincing argument that currency exposures would be
higher in an open economy. This may suggest that at least in the UK firms are controlling
for this exposure.



                Table 5 Stability of Firm Level Exposure across Sub Samples
The Table reports the number of firms for each exchange rate index that had significant
exposure at the 10% level for two or more sub periods. The average size of the sample
set was 419 firms


                                        Index              $         Yen        Euro
    Firms exposed in two             16 (3.8%)        11 (2.6%)   16 (3.8%)   19 (4.5%)
    periods
    Firms whose sign change               5                2          7          10
    between periods
    Firms exposed in all three       2 (0.05%)             0          0       2 (0.05%)
    sub periods
    Firms whose sign change                0               0          0           0



Sub Period Selection Model Results

The results of the selection model [3] are recorded in the lower section of Table 6 a-d.
Table 6a presents the exposure coefficients for the exchange rate index (TWI), 6b the
Dollar, 6c the Yen and 6d the Euro. Overall, there is an increase in the significance in the
models from period 1 to 3 for the exchange rate index and the Euro. The Yen model is not
significant for any period and the Dollar model is significant in period 1 and 2.

Where significant, the capital gearing ratio (CGR) is negative for all periods, firms with
high capital gearing ratios are not exposed to exchange rate risk indicating that firms are
managing potential financial distress using derivatives or other hedging vehicles. The
explanatory variables book to market rank (BMCAT) and the cash earnings per share
(CASHEPS) variables that measure underinvestment are positive and consistent with this
hypothesis. The positive result on the variable BMCAT suggests that exposed firms are
only those that do not have any significant growth potential and the positive result on
CASHEPS indicates that firms are less sensitive to the agency cost of debt. The results of
the selection model lend support to the hypothesis that firms are hedging their currency
exposure effectively and any residual exposures are strategic.


4
    Or taking advantage of any positive exchange rate movements
                                                        13
International operations (EXPORTS) are important determinants of exposure however
there are inconsistencies in the sign of the significant exposures. The evidence suggests
that in the earlier periods 1985-1989 and 1990-1994 there was a negative association
between exposure and international operations indicating that firms with low levels of
international operations were not able to control for the risks of international trade. In the
latter period, the association showed signs of a positive association between exposure and
international operations. These results suggest that a firm’s exposure to international
trade may be more complex than prior research evidence would suggest (Jorion, 1990)
and supports the argument of dynamic instability in exchange rates. This study supports
the argument there are cross sectional differences in exposure reflected in bilateral trade
in individual currencies, this together with cross sectional differences in currency volatility,
may result in the time variation in exposures. All models indicate that the larger the
organisation the greater the exposure to exchange rate risk. The results would suggest
that firms in the UK are effectively managing their exposures through hedging and
exposure is a function of size.


Sub Period Outcome Model Results

The results can be interpreted as identifying firms that are willing to carry exposure or are
unable to identify and hedge exposure. The results are recorded in the upper section of
Table 6.

The number of significant observations is low but this is consistent with prior research
findings (Jorian 1990). The results of the Heckman model supports the hypothesis that
firms are effectively hedging currency exposure and any residual exposures are accepted
as part of firms risk management strategy. For those firms that are exposed the beta
coefficient for BMCAT is positive and consistent with the argument that the only firms that
are exposed are those with low or poor growth opportunities. A positive and significant
value of the liquidity variable QAR indicates that firms are using liquidity as a hedging
substitute. Where significant the sign on the TAX variable is negative indicating that
increasing levels of tax is not a determinant of exposure and this is consistent for all
currency models. The inference from this suggests that at least in the UK firms are happy
to accept some level of exchange rate risk and this supports the prior survey findings of
Bradley and Moles (2001). They report that UK firms absorb adverse exchange rate
movements in order to protect market share. They also indicate that UK firms perceived
sensitivity to exchange rate exposure is low.

These findings indicate that financial distress (CGR) and the extent of international
operations (EXPORTS) is not a determinant of currency exposure using sub period
analysis.

There is also evidence to support significant shifts in the level of significance of the
explanatory variables between periods and currencies indicating that firm exposure varies
through time and is subject to firm specific shocks caused by economic trends.
Interestingly, although this study has reported that firm exposure was at its highest
during the period 1990-1994 the exposure model suggests UK firms were not exposed to
exchange rate risk during this period. As the Lambda is also insignificant, there is no
evidence to suggest that there is a missing variable problem influencing exposure that has
not been controlled for. Although there can be a number of explanations for these findings
to include effective exchange rate risk management during a shorter time period further
analysis on a wider set of variables may prove useful.




                                              14
                        Table 6 Heckman Two-Step Estimation

Table 6a Sub Period Selection/Outcome Model. A two step estimation for the
Exchange rate index (TWI) for three periods. The Selection Model represents a
probit model and first stage. The second stage is an Exposure Model where |βc0i|* was
included only if it was significant at 10%


                1985-89                   1990-94              1995-99
|βc0i|*         TWI                       TWI                  TWI
                Coef.          P >z       Coef.      P>z       Coef.       P>z
BMCAT              .06086         0.839   .05449       0.381      .05906     0.061
CGR               -.00113         0.967 -.00154        0.852     -.00013     0.979
TAX               -.03511         0.510   .00229       0.521     -.01212     0.028
QAR                .56990         0.726 -.91734        0.171      .16395     0.045
EXPORTS           -.01870         0.706 -.00418        0.548      .00612     0.126
CONSTANT         -5.44310         0.770 -1.25462       0.564     -.50499     0.641
Selection
Model
BMCAT              .03714         0.245   .02829       0.263   .03142        0.268
CGR               -.00009         0.974 -.00201        0.488 -.00683         0.063
CASH EPS          -.00135         0.667   .00172       0.431 -.00053         0.867
TAX                .00013         0.966   .00087       0.531 -.00127         0.679
QAR                .18628         0.270 -.34933        0.045   .08300        0.346
EXPORTS           -.00579         0.087 -.00287        0.301   .00517        0.070
LNMV               .02681         0.619   .05608       0.215   .12441        0.019
CONSTANT         -1.69855         0.022 -1.31157       0.042 -2.94175        0.000

LAMBDA            4.35630         0.674     2.1841     0.225     .91612      0.095
 No. of obs           397                      445                  416
 Censored Ops         348                      360                  364
 Uncensored            49                       85                   52
Ops
 Chi2 (4)               5.41                 13.61                25.28
 Prob >chi 7                     0.8619               0.1913                0.0048




                                                15
Table 6b Sub Period Selection/Outcome Model. A two step estimation for the US
Dollar (US$) for three periods

The Selection Model represents a probit model and first stage. The second stage is an
Exposure Model where |βc0i|* was included only if it was significant at 10%

                1985-89                1990-94               1995-99
|βc0i|*         US $                   US $                  US $
                Coef.       P >z       Coef.       P>z       Coef.       P>z
BMCAT             -.02225      0.306     .03990      0.022     -.03927     0.431
CGR                .00307      0.513     .00334      0.137      .00709     0.111
TAX               -.00340      0.492     .00048      0.574     -.00715     0.409
QAR               -.04472      0.677    -.17069      0.115     -.04676     0.769
EXPORTS           -.00372      0.144     .00014      0.940      .00555     0.510
CONSTANT           .08259      0.864    -.29533      0.541      2.3684     0.170
Selection
Model
BMCAT             -.02270      0.467   .04545        0.091   .04278        0.161
CGR               -.00971      0.095   .00004        0.973 -.00161         0.644
CASH EPS           .00048      0.623   .00159        0.477 -.00547         0.163
TAX               -.00008      0.988   .00105        0.422 -.00084        0.1810
QAR               -.08585      0.615 -.12862         0.415 -.03029         0.802
EXPORTS           -.00735      0.028 -.00309         0.271 -.00822         0.014
LNMV               0.9435      0.064   .13060        0.007   .04074        0.506
CONSTANT         -1.45876      0.048 -2.76672        0.000 -1.58977        0.050

LAMBDA             .54533      0.153     .50003      0.102     -.45328     0.650
 No. of obs           397                   445                    416
 Censored Ops         335                   378                    379
 Uncensored            62                    67                     37
Ops
 Chi2 (4)           17.03                 22.03                 14.89
 Prob >chi 7                  0.0736                0.0150                0.1360




                                              16
Table 6c Sub Period Selection/Outcome Model. A two step estimation for the
Yen for three periods
The Selection Model represents a probit model and first stage. The second stage is an
Exposure Model where |βc0i|* was included only if it was significant at 10%


                1985-89                   1990-94                 1995-99
|βc0i|*         Yen                       Yen                     Yen
                Coef.          P >z       Coef.         P>z       Coef.         P>z
BMCAT              .10198         0.484     .04306        0.101      .00850           0.484
CGR                .01274         0.472     .00002        0.989      .00047           0.823
TAX               -.00260         0.905     .00109        0.340     -.00243           0.141
QAR               -.01147         0.985    -.07875        0.500      .06268           0.561
EXPORTS            .00820         0.623     .00088        0.700     -.00152           0.294
CONSTANT          4.30724         0.393    -.47145        0.669      .03252           0.905
Selection
Model
BMCAT             -.03261         0.300   .05540          0.050     -.00838       0.754
CGR               -.00044         0.893 -.00017           0.909     -.00565       0.128
CASH EPS          -.00315         0.431   .00127          0.601      .00666       0.008
TAX               -.00048         0.890   .00146          0.309     -.00417       0.164
QAR                .06031         0.716   .09101          0.514     -.24443       0.178
EXPORTS           -.00351         0.272 -.00165           0.570     -.00152       0.591
LNMV              -.02303         0.664   .08079          0.109     -.02394       0.638
CONSTANT          -.44476         0.534 -2.58303          0.000     -.28642       0.683

LAMBDA            -2.9173         0.442     .56348        0.322     .40077        0.065
 No. of obs           397                      445                     416
 Censored Ops         337                      388                     349
 Uncensored            60                       57                      67
Ops
 Chi2 (4)               3.73                    10.60                   11.06
 Prob >chi 7                     0.9587                  0.3897                   0.3527




                                                17
Table 6d Sub Period Selection/Outcome Model. A two step estimation for the
Euro for three periods
The Selection Model represents a probit model and first stage. The second stage is an
Exposure Model where |βc0i|* was included only if it was significant at 10%


                1985-89                   1990-94               1995-99
|βc0i|*         Euro                      Euro                  Euro
                Coef.          P >z       Coef.       P>z       Coef.       P>z
BMCAT              .04363         0.812   .00389        0.970      .02372    0.370
CGR               -.00183         0.928 -.01768         0.501      .00166    0.776
TAX               -.02352         0.464   .00276        0.820     -.00219    0.413
QAR               -.08705         0.908 -1.17420        0.245      .12803    0.052
EXPORTS            .00248         0.812 -.00650         0.597      .00593    0.196
CONSTANT         -3.35900         0.819 -1.98320        0.614     -.48010    0.640
Selection
Model
BMCAT              .02544         0.417    -.00265      0.921   .01470       0.604
CGR               -.00072         0.812    -.00720      0.061 -.00997        0.018
CASH EPS          -.00114         0.656     .00295      0.193 -.00120        0.712
TAX               -.00229         0.569     .00024      0.887 -.00206        0.478
QAR               -.01792         0.920    -.23054      0.161   .01900       0.826
EXPORTS            .00006         0.984    -.00313      0.279   .00733       0.009
LNMV               .01791         0.729     .01316      0.782   .10358       0.049
CONSTANT          -1.3540         0.062    -.64965      0.330 -2.42011       0.001

LAMBDA             2.9988         0.738     3.4099      0.344     .80134      0.179
 No. of obs           397                      445                   416
 Censored Ops         345                      370                   360
 Uncensored            52                       75                    56
Ops
 Chi2 (4)               1.64                 10.35                 19.84
 Prob >chi 7                     0.9984                0.4104               0.0308




                                                 18
5. The Effect of Survivorship Bias

In previous studies there is a potential selection bias as the variable selection process has
therefore excluded part of the corporate population that may have been reporting data for
some but not all of the time frame in question. However there is a possibility that firm
attrition may have a significant affect on the overall results of prior studies. If firms were
excluded because they had not survived for the whole period, relevant information on
exchange rate exposure that contributes to financial distress would be lost.

Firms are recorded as ‘survivors’ if they survive from one period to the next. The periods
are identified in five-year blocks as before: 1985-89, 1990-94, 1995-99, and additionally,
2000-2004. To meet the survivorship criteria each firm needs to have reported continual
financial data available on DATASTREAM for the subsequent five-year period. Therefore for
a firm to qualify as a survivor for the period 1985 to 1989 it needs to report data for the
whole period 1990-1995. Data on survivors is collected for three groups corresponding to
the three sub periods and a final composite group that survived the entire 20 year period
1985 to 2004. Those that survive 1985-1989 are reported as group A those that survive
1990-1994 group B and those that survive 1995-99 group C. The final group D represents
those firms that survive the whole period 1985-2004.


                              Table 8 Survivor Group Samples
Group A represents those firms that report data in sub period 1 and 2. Group B in sub
periods 2 and 3, Group C in sub periods 3 and 4. Finally Group D represents those firms
that were reporting for the whole period


                         Group A          Group B             Group C          Group D
 Number of firms         397              445                 416
 Number selected         364              368                 282              190
 Number not selected     33               77                  134
 (non survivors)




Modelling Survivorship Bias

In order to fully investigate the possibility of survivorship or selection bias in the data set,
a Heckit survivorship model is used which incorporates two steps: the first is the selection
or survivor equation that identifies surviving firms (Survivor j), and the second step or
exposure model only includes those firms that survive. The selection equation is a
binomial probit estimation whereby the dependent variable is dichotomous and equal to 1
if the firm survives and is selected and 0 otherwise. In the selection equation variables are
used that may have a significant effect on the probability of the firm surviving is
consistent with prior empirical methods of failure prediction by Altman et al. (1993),
Curram et al. (1994), and Bell et al. (1990). That is those variables that model financial
distress, size, and short-term assets and liabilities:

                          Prob (Survivorj observed | gj) = Φ(gj γ)

where gi represents those variables that determine survival.

                                                19
The Heckit generated variable lambda (λ), if significant, indicates that there is a missing
variable problem and survivorship bias and makes corrections within the second step of
the model. The second step or exposure model only includes those variables that meet the
survival criteria. In the second step, the dependent variable is defined as the absolute
value of beta on the exchange rate index or bilateral exchange rate as before from [1],
and the explanatory variables are those variables that determine exposure:

                                                          /
                          E[|βc0i| | Survivor j =1] = β       zj + θλ   j


where zi represents those variables that determine exposure as before. The survival issue
is investigated by running the model on groups A-D.


Tests on Full Period Survival

Firstly, in order to make a representative comparison with prior research, a data set is
constructed to include only those firms that survived the whole period 1985-2004 (Group
D). This group is similar to the sample used by Jorion (1990) and Choi and Prasad (1995).
Firstly, univariate tests are run to determine whether survivorship displays any significant
difference from firms that survive the whole period to those firms that did not survive. The
analysis is conducted for each of the three sub periods. Surviving firms are recorded as 1
indicating that they survive each of the three sub periods and up to December 2004. The
firms that do not survive are recorded as 0. The univariate results are recorded in Table 9
and suggest that surviving firms are larger and have a higher level of international
operations than non-surviving firms. This is consistent across all three time periods.

Secondly, multivariate analysis is undertaken using the Heckit two step model (see Table
10) as a test for survivorship bias over the three sub periods.

The results indicate that lambda (λ) is significant for the Yen in sub period 1 indicating
that survivorship bias has been detected and may be an issue influencing the results. The
positive significant coefficient can be interpreted as indicating if more firms survive this
would have a positive impact on exposure. The multivariate results indicate that surviving
firms are larger and have a higher level of international operations than non surviving
firms. There is also evidence that internally generated funds are an important determinant
of survival. This is consistent across all three time periods. For those firms that do survive
however, the significance of the second step model that identifies exposure determinants
although small varies across the bilateral exchange rates. The majority of prior studies
only used a composite exchange rate index in their sample and the results of this study
reveals no significant exposure determinants in surviving firms, indicating that exposure to
a trade weighted index is not influenced by the determinants as predicted by financial
theory This may suggest that the use of a trade weighted index may not be a reliable in
detecting currency exposure. These results are consistent with those of Jorion (1990) and
Bartov and Bodnar (1994), indicating that US firms are not exposed to exchange rate risk
using a currency index.




                                             20
Table 9a Univariate Analysis, Survivors and Non Survivors
  Group Statistics                                                          t-test for   Equality   of
  1985-89                                                                   Means
                                                               Std.                      Sig.    (2-
                            Survivors=1   N        Mean        Deviation    T            tailed)
  TWI                       1             190      0.01289     0.49921      1.50275      0.1337
                            0             207      -0.06781    0.56498      1.51072      0.1316
  US$                       1             190      0.05312     0.33973      2.75427      0.0061
                            0             207      -0.04127    0.34239      2.75519      0.0061
  EURO                      1             190      0.00198     0.49812      0.20122      0.8406
                            0             207      -0.00858    0.54356      0.20197      0.8400
  YEN                       1             190      -0.08679    0.47062      1.10155      0.2713
                            0             207      -0.13861    0.46610      1.10109      0.2715
  BMCAT                     1             190      6.09473     2.653033     -0.36780     0.7132
                            0             207      6.19323     2.677000     -0.36794     0.7131
  CGR                       1             190      28.85121    44.037000    1.08382      0.2791
                            0             207      25.32271    15.303022    1.04791      0.2957
  CASHEPS                   1             190      22.38913    31.779211    -0.46075     0.6452
                            0             207      26.05595    105.379700   -0.47753     0.6334
  TAX                       1             190      35.37341    36.253490    2.14055      0.0329
                            0             207      29.43262    16.041622    2.07961      0.0385
  QAR                       1             190      0.95803     0.469682     -0.85732     0.3917
                            0             207      1.00018     0.506700     -0.86011     0.3902
  EXPORTS                   1             190      31.69455    28.123100    1.68423      0.0929
                            0             207      27.08377    26.419350    1.67973      0.0938
  LNMV                      1             190      11.76634    1.685510     3.18041      0.0015
                            0             207      11.25613    1.510710     3.16556      0.0016




                                                              21
Table 9b Univariate Analysis, Survivors and Non Survivors
                                                                           t-test for   Equality   of
  Group Statistics 1999-1994                                               Means
                                                               Std.                     Sig.    (2-
                               Survivors=1   N     Mean        Deviation   T            tailed)
  TWI                          1             190   0.27769     0.75104     0.37476      0.7080
                               0             255   0.25013     0.77936     0.37679      0.7065
  US$                          1             190   0.00063     0.38966     -0.32548     0.7449
                               0             255   0.01346     0.42685     -0.32984     0.7416
  EURO                         1             190   0.39850     0.83805     0.78135      0.4350
                               0             255   0.33659     0.81824     0.77862      0.4366
  YEN                          1             190   0.03612     0.39140     -0.01124     0.9910
                               0             255   0.03657     0.44111     -0.01143     0.9908
  BMCAT                        1             190   6.35789     2.94361     0.53140      0.5954
                               0             255   6.20784     2.94836     0.53152      0.5953
  CGR                          1             190   34.81637    72.63795    0.85052      0.3954
                               0             255   30.63946    25.56797    0.75839      0.4490
  CASHEPS                      1             190   24.68150    24.38346    0.96921      0.3329
                               0             255   21.70013    36.80167    1.02620      0.3053
  TAX                          1             190   29.97773    27.68269    -0.53502     0.5929
                               0             255   32.47086    59.60879    -0.58817     0.5567
  QAR                          1             190   1.02028     0.55496     0.12545      0.9002
                               0             255   1.01346     0.57623     0.12614      0.8996
  EXPORTS                      1             190   35.75644    29.02026    1.75893      0.0792
                               0             255   30.85805    29.08667    1.75952      0.0792
  LNMV                         1             190   12.26948    1.71928     2.74174      0.0063
                               0             255   11.82743    1.65429     2.72633      0.00668




                                                              22
Table 9c
                                                                           t-test for   Equality   of
  Group Statistics 1995-1999                                               Means
                                                               Std.                     Sig.    (2-
                               Survivors=1   N     Mean        Deviation   T            tailed)
  TWI                          1             190   -0.12411    0.72461     -1.06442     0.2877
                               0             226   -0.04642    0.75548     -1.06827     0.2860
  US$                          1             190   -0.08731    0.63898     0.76520      0.4445
                               0             226   -0.14168    0.78482     0.77878      0.4365
  EURO                         1             190   -0.18029    0.66975     -1.10209     0.2710
                               0             226   -0.10510    0.71214     -1.10796     0.2685
  YEN                          1             190   0.11506     0.33339     -1.24587     0.2135
                               0             226   0.15611     0.33594     -1.24669     0.2132
  BMCAT                        1             190   5.82105     3.17223     0.98664      0.3243
                               0             226   5.51327     3.16681     0.98649      0.3244
  CGR                          1             190   52.96789    240.9525    1.25272      0.2110
                               0             226   32.75428    26.10052    1.15068      0.2512
  CASHEPS                      1             190   34.04675    27.29260    -0.91187     0.3623
                               0             226   1133.410    16615.060   -0.99470     0.3209
  TAX                          1             190   29.04876    16.54803    1.34657      0.1788
                               0             226   25.34622    34.72591    1.42226      0.1558
  QAR                          1             190   1.01702     0.64794     0.11018      0.9123
                               0             226   1.00799     0.95958     0.11381      0.9094
  EXPORTS                      1             190   41.03394    30.63533    2.39870      0.0168
                               0             225   33.84900    30.20227    2.39581      0.0170
  LNMV                         1             190   12.84453    1.70675     2.11934      0.0346
                               0             226   12.4892     1.70049     2.11867      0.0347




                                                              23
           Table 10. Heckit 2 Step Model. Survivor for the Whole Period

Table 10a. Exposure Coefficients for the Exchange Rate Index TWI. All Firms
listed on the UK All Share Index. Significant exposures at the 10% are in bold. A survivor
is a firm that survives from 1985 to end 2004. Survivor whole period represents a
probit model and first stage. |βc0i| the second stage. The explanatory variables were
calculated from an average of the five year period.

                1985-89                1990-94                1995-99
|βc0i|          TWI                    TWI                    TWI

                Coef.       P >z       Coef.        P>z       Coef.       P>z
BMCAT             .001135      0.907 0.01852          0.137     0.01257     0.318
CGR               .000755      0.226 -0.00013         0.803    -7.9E-05     0.674
TAX              -.000527      0.418 0.00019          0.878    -0.00336     0.101
QAR               .068665      0.262 -0.07753         0.237    -0.08332     0.146
EXPORTS          -.000741      0.501 -0.00061         0.682     0.00302     0.113
CONSTANT          .169343      0.333 0.37573          0.270     0.25564     0.547
Survivor
whole period
BMCAT             .002236      0.929 0.03033          0.171     0.02595    0.215
CGR               .001782      0.428 0.00057          0.673     0.00057    0.517
CASH EPS         -.000760      0.480 0.00029          0.879     0.00382    0.096
QAR              -.100169      0.475 0.02029          0.857     0.01241    0.874
EXPORTS           .002475      0.331 0.00251          0.259     0.00482    0.025
LNMV              .123545      0.004 0.10709          0.007     0.05772      0.15
CONSTANT         -1.48626      0.010 -1.78763         0.001    -1.31294    0.015

LAMBDA            .193127      0.342    0.14054       0.647     0.35696         0.34
 No. of obs           397                   445                     416
 Censored obs         207                   255                     226
 Uncensored           190                   190                     190
Ops (9)
 Chi2                5.78                  9.04                   15.28
 Prob >chi 2                  0.7622                 0.4332               0.0836




                                               24
Table 10b Exposure Coefficients for the Dollar. All Firms listed on the UK All Share
Index. Significant exposures at the 10% are in bold. A survivor is a firm that survives
from 1985 to end 2004. Survivor whole period represents a probit model and first
stage; |βc0i| the second stage. The explanatory variables were calculated from an
average of the five year period

                1985-89                   1990-94             1995-99
|βc0i|          US $                      US $                US $

                Coef.      P >z           Coef.     P>z       Coef.      P>z
BMCAT            0.00455          0.468   0.00278     0.679    0.01213         0.088
CGR              0.00040          0.300   0.00048     0.084    0.00019         0.576
TAX             -0.00060          0.169         -     0.407    0.00171         0.095
QAR              0.01020          0.797   0.00055
                                                -     0.285    0.05218         0.307
EXPORTS         -0.00074          0.293   0.03774
                                                -     0.747    0.00180         0.713
CONSTANT         0.31049          0.006   0.00026
                                          0.37224     0.043    0.40228         0.878
Survivor
whole period
BMCAT            0.00223          0.929   0.03033     0.171    0.02091         0.215
CGR              0.00178          0.428   0.00057     0.673    0.00089         0.517
CASH EPS        -0.00076          0.480   0.00029     0.879    0.00229         0.096
QAR             -0.10016          0.475   0.02029     0.857    0.07858         0.874
EXPORTS          0.00247          0.331   0.00251     0.259    0.00215         0.025
LNMV             0.12354          0.004   0.10709     0.007    0.04009           0.15
CONSTANT        -1.48626          0.010         -     0.001    0.54114         0.015
CONSTANT                                  1.78763
LAMBDA          -0.06183          0.636         -     0.568    0.35601         0.135
 No. of obs          397                  0.09448
                                              445                  416
 Censored obs        207                      255                  226
 Uncensored          190                      190                  190
Ops (9)
 Chi2               7.43                     9.04                15.16
 Prob >chi 2      0.5929                             0.4332                0.0867




                                              25
Table 10c Exposure Coefficients for the Yen. All Firms listed on the UK All Share
Index. Significant exposures at the 10% are in bold. A survivor is a firm that survives
from 1985 to end 2004. Survivor whole period represents a probit model and first
stage; |βc0i| the second stage. The explanatory variables were calculated from an
average of the five year period

               1985-89                     1990-94                   1995-99
|βc0i|         Yen                         Yen                       Yen

               Coef.        P >z           Coef.      P>z            Coef.      P>z
BMCAT            -0.00103          0.939 0.00874            0.223 0.00016              0.979
CGR               0.00105          0.280 0.00049            0.091 -4.12E-06            0.966
TAX              -0.00088          0.211 7.83E-06           0.991 0.00042              0.649
QAR              -0.01546          0.853 -0.00918           0.808 0.00787              0.774
EXPORTS          -0.00025          0.867 0.00074            0.385 9.01E-05             0.923
CONSTANT         -0.09865          0.694 0.15593            0.425 0.05659              0.786
Survivor
whole
period
BMCAT             0.00223          0.929 0.03033            0.171 0.02595             0.215
CGR               0.00178          0.428 0.00057            0.673 0.00057             0.517
CASH EPS         -0.00076          0.480 0.00029            0.879 0.00382             0.096
QAR              -0.10016          0.475 0.02029            0.857 0.01241             0.874
EXPORTS           0.00247          0.331 0.00251            0.259 0.00482             0.025
LNMV              0.12354          0.004 0.10709            0.007 0.05772               0.15
CONSTANT         -1.48626          0.010 -1.78763           0.001 -1.31294            0.015

LAMBDA            0.56970          0.048    0.03843          0.828    0.23961          0.193
 No. of obs           397                       445                       416
 Censored             207                       255                       226
obs
 Uncensored           190                       190                       190
Ops (9)
 Chi2                5.16                      8.88                      7.89
 Prob >chi 2       0.8198                                   0.4480                    0.5450




                                                 26
Table 10d Exposure Coefficients for the Euro. All Firms listed on the UK All Share
Index. Significant exposures at the 10% are in bold. A survivor is a firm that survives
from 1985 to end 2004. Survivor whole period represents a probit model and first
stage; |βc0i| the second stage. The explanatory variables were calculated from an
average of the five year period

               1985-89                     1990-94                  1995-99
|βc0i|         Euro                        Euro                     Euro

               Coef.        P >z           Coef.      P>z           Coef.        P>z
BMCAT             0.00364          0.739 0.03026        0.088          0.00559          0.624
CGR               0.00041          0.581 0.00093        0.249         -6.8E-05          0.676
TAX              -0.00042          0.523 0.00253        0.093         -0.00311          0.111
QAR               0.03907          0.569 -0.03478       0.708         -0.06726          0.205
EXPORTS           0.00069          0.576 0.00016        0.937          0.00348          0.045
CONSTANT          0.00918          0.903 -0.29966       0.539          0.35093          0.364
Survivor
whole
period
BMCAT             0.00223          0.929 0.03033        0.171          0.02595          0.215
CGR               0.00178          0.428 0.00057        0.673          0.00057          0.517
CASH EPS         -0.00076          0.480 0.00029        0.879          0.00382          0.096
QAR              -0.10016          0.475 0.02029        0.857          0.01241          0.874
EXPORTS           0.00247          0.331 0.00251        0.259          0.00482          0.025
LNMV              0.12354          0.004 0.10709        0.007          0.05772            0.15
CONSTANT         -1.48626          0.010 -1.78763       0.001         -1.31294          0.015

LAMBDA            0.34759          0.134    0.68728         0.115      0.19662          0.563
 No. of obs           397                       445                        416
 Censored             207                       255                        226
obs
 Uncensored           190                       190                        190
Ops (9)
 Chi2                3.98                     12.03                      16.18
 Prob >chi 2                   0.9127                   0.2115                         0.0632




                                               27
Tests on Period to Period Survival

The Heckit selection model was again used to determine survival based on groups A, B,
and C as defined in Table 8. In this instance, the period of study for investigation is the
sub periods (1), (2) and (3). The results reported in table 11 indicate that survivorship is
largely determined by size, as financial theory would predict; the larger the firm the
greater the chance of survival. During period two (1990-1994) there is evidence of
significant survivorship bias across all four currency variables. The Lambda is significant
and positive indicating that if firms survived this would have the effect of increasing firm
exposure to exchange rate risk. As this period represented a period of UK economic
instability and exchange rate volatility the vulnerability of UK firms would have been
significantly higher. There is therefore strong evidence to suggest that survivorship bias
may question the reliability of exchange rate risk exposure studies especially in times of
economic instability. The decomposition of the data set into sub periods shows that in
times of economic uncertainty the level of firm exposure rises and therefore this may
have the effect of increasing corporate failure. If this were the case then previous studies
of exchange rate exposure identification would have underestimated the extent of
exposure.


For those firms that do survive and are exposed to exchange rate risk the results are
consistent with financial theory and earlier results, the BMCAT and CGR is positive
indicating that firms high book to market ranks and high gearing ratios are exposed to
exchange rate risk. Again, and consistent with prior results, there is a positive association
between the variable EXPORTS and the Euro beta.




                                             28
                           Table 11 Heckit Selection Model

Table 11a Survivor Model Exposure Coefficients for the Exchange rate index
TWI. All Firms listed on the UK All Share Index. Significant exposures at the 10% are in
bold. A survivor is a firm that survives to the end of the next period t+1. Survivor
represents a probit model and first stage. |βc0i| represents the second stage on a
restricted model where the absolute value of βc0 was included if the firms survived. The
explanatory variables were calculated from an average of the five year period.

                1985-89                   1990-94               1995-99
                Period 1                  Period 2              Period 3
|βc0i|          TWI                       TWI                   TWI

                Coef.       P >z      Coef.     P>z     Coef.     P>z
BMCAT              -.028702     0.786   .022119   0.159   .001928   0.867
CGR                -.001841     0.809 -.000024    0.975 -.000075    0.698
TAX                -.000704     0.897 -.000906    0.273 -.001708    0.121
QAR                 .000035     1.000 -.070947    0.333 -.091714    0.088
EXPORTS            -.003425     0.718 -.001737    0.228   .001484   0.179
CONSTANT            .145774     0.849   .328766 0.032     .360680   0.088
Survivor
BMCAT               -.032287      0.361    -.016173    0.533     -.011145   0.603
CGR                 -.001434      0.517    -.000938    0.443      .000294   0.647
CASH EPS             .000615      0.787    -.001168    0.605      .001164   0.636
QAR                 -.007879      0.971     .040853    0.774     -.062870   0.424
EXPORTS             -.004052      0.252    -.002151    0.420     -.000426   0.848
LNMV                 .051221      0.407     .217281    0.000      .084711   0.045
CONSTANT            1.165610      0.162    -1.41892    0.045     -.507889   0.368

LAMBDA              3.601240      0.642     .807463    0.019      .571112    0.193
 No. of obs              397                    445                   416
 Censored obs               33                   77                   134
 Uncensored                364                  368                   282
obs
 Chi2 (9)                3.31                   9.91                 7.05
Prob >chi 2                      0.9505                0.3578               0.6321




                                           29
Table 11b Survivor Model Exposure Coefficients for the Dollar. All Firms listed on
the UK All Share Index. Significant exposures at the 10% are in bold. A survivor is a firm
that survives to the end of the next period t+1. Survivor represents a probit model and
first stage. |βc0i| represents the second stage on a restricted model where the absolute
value of βc0 was included if the firms survived. The explanatory variables were calculated
from an average of the five year period.



                 1985-89                   1990-94                        1995-99
                 Period 1                  Period 2                       Period 3
  |βc0i|         US $                      US $                           US $

                 Coef.          P >z       Coef.           P>z            Coef.          P>z
  BMCAT             -.007346       0.828      .010039            0.098     -.005140            0.774
  CGR               -.000291       0.904      .000684            0.022      .000129            0.672
  TAX               -.000619       0.722     -.000547            0.108     -.003050            0.060
  QAR               -.008640       0.948     -.012954            0.645     -.024389            0.767
  EXPORTS           -.001506       0.618     -.000897            0.106     -.000320            0.851
  CONSTANT           .201804       0.409      .166280            0.005     -.000320            0.529
  Survivor
  BMCAT             -.032287       0.361     -.016173            0.533     -.011145        0.603
  CGR               -.001434       0.517     -.000938            0.443      .000294        0.647
  CASH EPS           .000615       0.787     -.001168            0.605      .001164        0.636
  QAR               -.007879       0.971      .040853            0.774     -.062870        0.424
  EXPORTS           -.004052       0.252     -.002151            0.420     -.000426        0.848
  LNMV               .051221       0.407      .217281            0.000      .084711        0.045
  CONSTANT          1.165616       0.162    -1.418924            0.045     -.507889        0.368

  LAMBDA            1.148786       0.642      .257702            0.053     .952236             0.162
   No. of obs            397                      445                          416
  Censored Obs            33                       77                          134
  Uncensored             364                      368                          282
  Obs
   Chi2 (9)              3.55                      17.54                          5.06
   Prob >chi 2                    0.9384                         0.0410                    0.8294




                                             30
Table 11c. Survivor Model Exposure Coefficients for the Yen. All Firms listed on the
UK All Share Index. Significant exposures at the 10% are in bold. A survivor is a firm that
survives to the end of the next period t+1. Survivor represents a probit model and first
stage. |βc0i| the second stage on a restricted model where the absolute value of βc0 was
included if the firms survived. The explanatory variables were calculated from an average
of the five year period.

                  1985-89                       1990-94                 1995-99
                  Period 1                      Period 2                Period 3
|βc0i|            Yen                           Yen                     Yen

                  Coef.          P >z           Coef.        P>z        Coef.          P>z
BMCAT               -.026315            0.800 .011301              0.091 -.006243             0.420
CGR                 -.001734            0.816 .000552              0.099 3.22e-06             0.981
TAX                 -.001648            0.758 -.000476             0.177 .000378              0.589
QAR                  .011971            0.977 .016554              0.596 .002237              0.950
EXPORTS             -.003657            0.694 -.000295             0.631 -.000357             0.629
CONSTANT             .110480            0.883 .106675              0.103 .100531              0.478
Survivor
BMCAT               -.032287            0.361   -.016173           0.533    -.011145         0.603
CGR                 -.001434            0.517   -.000938           0.443     .000294         0.647
CASH EPS             .000615            0.787   -.001168           0.605     .001164         0.636
QAR                 -.007879            0.971    .040853           0.774    -.062870         0.424
EXPORTS             -.004052            0.252   -.002151           0.420    -.000426         0.848
LNMV                 .051221            0.407    .217281           0.000     .084711         0.045
CONSTANT            1.165616            0.162   -1.41892           0.045    -.507889         0.368

LAMBDA              3.528112            0.642    .344174           0.019    .411847           0.162
No. of obs               397                         445                        416
Censored Ops              33                          77                        134
Uncensored obs           364                         368                        282
Chi2 (9)                  3.42                       10.55                      2.28
 Prob >chi 2                        0.9454                         0.3079                    0.9063




                                                31
Table11d. Survivor Model Exposure Coefficients for the Euro. All Firms listed on the
UK All Share Index. Significant exposures at the 10% are in bold. A survivor is a firm that
survives to the end of the next period t+1. Survivor represents a probit model and first
stage. |βc0i| represents the second stage on a restricted model where the absolute value of
βc0 was included if the firms survived. The explanatory variables were calculated from an
average of the five year period.

                  1985-89                      1990-94                     1995-99
                  Period 1                     Period 2                    Period 3
|βc0i|            Euro                         Euro                        Euro

                  Coef.         P >z           Coef.        P>z            Coef.      P>z
BMCAT                -.023574          0.811      .030658          0.108      .0002396      0.981
CGR                  -.001897          0.788      .000843          0.377     -.0000784      0.624
TAX                  -.000867          0.864     -.000333          0.740     -.0022295      0.042
QAR                   .001285          0.997     -.092135          0.303     -.0691787      0.154
EXPORTS              -.002939          0.718     -.001908          0.278      .0021981      0.020
CONSTANT              .137502          0.849       276199          0.138      .4320485      0.017
Survivor
BMCAT               -.032287           0.361     -.016173         0.533      -.0111455      0.603
CGR                 -.001434           0.517     -.000938         0.443       .0002949      0.647
CASH EPS             .000615           0.787     -.001168         0.605       .0011643      0.636
QAR                 -.007879           0.971      .040853         0.774      -.0628702      0.424
EXPORTS             -.004052           0.252     -.002151         0.420      -.0004261      0.848
LNMV                 .051221           0.407      .217281         0.000       .0847115      0.045
CONSTANT            1.165616           0.162    -1.418924         0.045      -.5078897      0.368
LAMBDA              3.344851           0.642      .979777         0.019       .3265393      0.387
 No. of obs              397                          445                          416
Censored Obs              33                           77                          134
Uncensored obs           364                          368                          282
 Chi2 (9)               3.32                        10.08                        10.37
 Prob >chi 2                       0.9504                         0.3439                    0.3213




                                               32
6. Conclusions

This paper examines the existence of a contemporaneous relationship between the stock
returns of UK non-financial firms and fluctuations in trade weighted and bilateral foreign
exchange rates using three sub periods. The sample was split into three equal periods
covering the years 1985-1989, 1990-1994, and 1995-1999. The evidence to suggest that
macro economic shocks are an important factor in the determination of the sensitivity of
share value to exchange rate volatility. These macro economic shocks were at their
greatest during the period 1990-1994 when the Pound depreciated against all major
currencies and the significant exchange rate exposure co-efficients were positive. These
results suggest that UK companies are at their greatest risk as the Pound depreciates
indicating that UK industry is highly sensitive to costs.


Using a Heckman two stage sub period selection/exposure model, tests were undertaken
to identify the significance of exposure determinants over three periods using trade
weighted and bilateral exchange rate indices. The selection model reports the
effectiveness of firms hedging strategy and the results suggest most UK firms are
effectively managing their exposure to exchange rate risk through hedging or utilising
hedging substitutes. Larger firms are prepared to carry more exposure than their smaller
counterparts. The exposure model reports on a firm’s ability to manage residual exchange
rate exposure and the results suggest UK firms are not exposed. Even when considering
data on three bilateral exchange rates, Dollar Yen and Euro the overall significance of the
model is low. This confirms that currency exposure in the UK does not seem to be
affected by the hypothesised variables deemed important in the determination of
exposure. Even the level of international activity is of no relevance.

These conclusions suggest that the insignificance of the sensitivity of firms’ value to
currency exposure reported by Jorion (1990) and Choi and Prasad (1996) was after all
correctly specified and consistent with the hypothesis reported by Bartov and Bodnar
(1994) that the insignificant results are a result of effective hedging. Where significant
exposure is detected it can be regarded as residual and voluntary.

Conventional models for constructing exposure determinants rest on the assumption that
the firms in the data sets are representatives of the business world, especially in
international operations. Unfortunately as firms can be omitted from the data set as they
have ceased trading, this creates a selection problem and survivorship bias. The
survivorship bias would be a serious problem if the significance of the model is
underestimated because of the elimination of high risk firms from the sample. In order to
test for survivorship bias, a Heckit selection model was used. The results suggest there is
strong evidence to support the hypothesis that UK firms are more likely to survive if they
are larger in size. Survivorship bias was detected and significant during the period 1990-
1994 when the UK exchange rates and interest rates were volatile and as a result could
have major implications for currency risk studies if the data set constitutes large firms
only as this is more likely to have a downward bias on detecting exposure.

In conclusion this paper provides evidence of significant time variation in exchange rate
exposure and reports exposure is at its highest at times of increased currency volatility.
There may also be a potential downward bias in the significance of exposure betas as firm
survivorship may be a determinant of exposure.




                                            33
References

Allayannis, G., and Ofek, E., (2001). Exchange rate exposure, Hedging and the use of
currency derivatives. Journal of International Money and Finance, March, 20, 2, 273-96.

Altman E.I., Marco G., and Varetto F., (1993). Corporated distress diagnosis:
comparisons using linear discriminant analysis and neural networks (the Italian
experience), Journal of Banking and Finance 18, pp. 505-529.

Amihud, Y., (1994). Exchange Rates and the Valuation of Equity Shares, in Amihud, Y. and
Levich, R., (ed), Exchange Rates and Corporate Performance, Irvin Professional Publishing.

Bartov, E., and Bodnar, G. M., (1994). Firm valuations, earnings expectations, and the
exchange rate exposure effect, Journal of International Business Studies. 24, 557-573.

Bartram, S,M., and Bodnar, G,M., (2007) The Foreign Exchange Exposure Puzzle,
Managerial Finance, Vol. 33, 9, 642-666.

Belk, P.A., and Glaum, M., (1990). The management of foreign exchange risk in UK
multinationals: An empirical investigation, Accounting and Business Research, 21, 3-13.

Belk, P., Bakey, V., and Duangploy, O., (1993). Foreign exchange risk management
practices: A comparative empirical analysis of UK and US MNE’s. Loughborough University
Business School, Working Paper.

Bell T., Ribar G., and Verchio, J., (1990). Neural nets vs. logistic regression: a
comparison of each model’s ability to predict commercial bank failures, Cash Flow
Accounting Conference, Nice (France), 18 pp.

Berkman, H., and Bradbury, M. E., (1996). Empirical evidence on the corporate use of
derivatives, Financial Management, 25, 2, 5-13.

Berkman, H., Bradbury, M. E., and Magan, S., (1997). An international comparison of
derivatives use. Financial Management, 26, 4, 69-73.

Bodnar, G.M., and Gentry, W.M., (1993). Exchange rate exposure and industry
characteristics: Evidence from Canada, Japan and the USA. Journal of International Money
and Finance, 12, 29-45.

Bodnar, G., and Wong, M., (2003). Estimating exchange rate exposure: issues in model
structure, Financial Management. 32 1, 35–67.

Booth, L., and Rotenberg, W., (1990). Assessing foreign exchange exposure; Theory and
application using Canadian firms. Journal of International Financial Management and
Accounting, 2, 1, 1-22.

Bradley, K.D., and Moles, P., (2001), The effect of exchange rate movements on non-
financial UK firms, International Business Review, 10, 51-69.

Brown, S. J., and Goetzmann, W., (1995). Performance persistence, Journal of Finance
50, 679–698.

Choi, J., and Prasad, A., (1995). Exchange risk sensitivity and its determinants: A firm and
Industry analysis of US multinationals, Financial Management, 24, 3, 77-88.

Chow, E.H., Lee, W.Y., and Solt, M.E., (1997). The economic exposure of U.S. Multinational


                                            34
Firms, The Journal of Financial Research, 20, 2, 191-210.

Curram, S., and Mingers, J., (1994). Neural Networks, Decision tree induction and
discriminant analysis: an empirical comparison, Operational Research Society, pp. 440-
450.

Davis, J. L., (1996). The cross-section of stock returns and survivor bias: Evidence from
delisted stocks. Quarterly Review of Economics and Finance 36, 365-375.

De Jong, A., Ligterink, J. E, and Macrae, V., (2002), A Firm-Specific Analysis of the
Exchange-Rate Exposure of Dutch Firms, ERIM Report Series Reference No. ERS-2002-
109-F&A.

Doidge, C., Griffin, J., and Williamson R., (2002). Does exchange rate exposure matter?
Working paper http://ssrn.com/abstract=313060.

Dominguez, K., and Tesar. L., (2001). A re-examination of exchange rate exposure,
American Economic Review papers and proceedings, May.

Donnelly, R., and Sheehy, E., (1996). The share price reaction of UK Exporters to exchange
rate movements: An empirical study, Journal of International Business Studies. 27, 157-
165.

Dunne, T. M., and Helliar, C., (2003). FRS 13: Implementation and disclosures,
Accountancy, June, 95-96.

Dunne,T.M., Hellier, C., Power, D., Mallin, C. A., Ow-Yong, K.H., and Moir.L., (2004). The
introduction of Derivatives Reporting in the UK: A Content Analysis of FRS13 Disclosures,
Journal of Derivatives Accounting, 1, 2, 205-219.

El-Masry, A. (2007), Foreign exchange rate exposure and its determinants of UK non-
financial companies, Managerial Finance, Vol. 33, No. 9, September. Forthcoming

Froot, K., Scharfstein, D.S., and Stein J.C., (1993). Risk management: Co-ordinating
corporate investment and financing policies. Journal of Finance, 48, 5, 1629-1658.

Gao, T., (2000). Exchange rate movements and the profitability of US multinationals.
Journal of international Money and Finance, 19, 117-134.

Geczy, C., Minton, B.A., and Schrand, C., (1997). Why firms use currency derivatives,
Journal of Finance, 52, 4, 1323-1353.

Glaum, M., Brunner, M., and Himmel, H., (1999). The DAX and the dollar: The economic
exchange rate exposure of German corporations , Journal of International Business Studies,
31, 4, 715-724.

Graham, J. R., and Rogers, D. A., (2002). Do firms hedge in response to tax incentives?,
Journal of Finance 57, 815-839.

Griffin, J., and Stulz, R., (2001). International competition and exchange rate shocks; A
cross-country industry analysis and stock returns, Review of Financial Studies 14, 215-
241
He, J., and Ng, L., (1998). The foreign exchange exposure of Japanese multinational
corporations, Journal of Finance, 53, 2, 733-741.




                                             35
Heckman, J., (1979). Sample Selection Bias as a Specification Error. E conometrica, 47(1):
153-161.

Howton, S. D., and Perfect, S. P., (1998). ‘Currency and interest-rate derivatives use in
US firms’. Financial Management, 27, 4, 111-122.

Ihrig, J., and Prior, D., (2004). The effect of exchange rate fluctuations on Multinational
returns. International finance Discussion Papers, International Finance Division, Board of
Governors of the Federal Reserve System,

Jorion, P., (1990). Exchange rate exposure of USA multinationals, Journal of Business, Oct,
331-345.

Joseph, N. L., (2000). The choice of hedging techniques and the characteristics of U.K.
industrial firms, Journal of Multinational Financial Management, 10,161-184

Kiymaz, H., (2003). Estimation of Foreign exchange exposure: an emerging market
application. Journal of multinational Financial Management. 13, 71-84

Kothari, S.P., Shanken, J., and Sloan, R.G., (1995). Another look at the cross-section of
expected stock returns. Journal of Finance 50, 185-224

Maker, S.D., DeBruin, J., and Huffman, S. P., (1999). The management of foreign currency
risk: Derivative use and the natural hedge of geographic diversification, Accounting and
Business Research, 29, 3, 229-237.

Maker, S.D., and Huffman, S. P., (2001). Foreign exchange derivatives, exchange rate
changes and the value of the firm: US multinationals’ use of short-term financial
instruments to manage currency risk, Journal of Economic and Business, 53, 421-437.

Martin, A.D., Madura, J., and Akhigbe, A., (1999). Economic exchange rate exposure of US
based MNC’s operating in Europe, The Financial review, 34, 21-36.

Miller, K., and Reuer, J., (1998). Firm strategy and economic exposure to foreign
exchange rate movements, Journal of International Business Studies, 29, 3, 493-514.

Nance, D.R., Smith, C.W., and Smithson, C.W., (1993). On the determinants of corporate
hedging, Journal Of Finance, 48, 267-284.

Nguyen, H. & Faff, R. (2002), On the determinants of derivative usage by Australian
companies, Australian Journal of Management, 27, 1, 1–24.

Nydahl, S., (1999). Exchange rate exposure. Foreign involvement in currency hedging of
firms-some Swedish evidence, Working paper, Quantal International.

Parsley, D, C., and Popper, H., Exchange Rate Pegs and Foreign Exchange Exposure in
East and South East Asia, (2006) , Journal of International Money and Finance, 25, 992-
1009.

Prasad, A.M., and Rajan, M., (1995). The role of exchange and interest risk in equity
valuation: A comparative study of international stock markets, Journal of Economics and
Business, 47, 457-472.

Salifu, Z., Osei, K, A., and Adjasi, C, K.D. (2007) Foreign exchange risk exposure of
listed companies in Ghana The Journal of Risk Finance Incorporating Balance Sheet, 8, 4,
380-393




                                            36
Shin, H. H., and Soenen, L., (1999). Exposure to currency risk by US multinational
corporations. Journal of Multinational Financial Management. 9,195-207

Smith, C. W., Smithson, C.W., and D, Sykes Wilford., (1990). Strategic Risk Management,
Institutional Investor Series in Finance, Harper and Row, New York.

Smith, C. W., and Stulz, R., (1985). The determinants of firms' hedging policies, Journal of
Financial and Quantitative Analysis, 20, 391-405,

Tufano, P., (1996). Who manages risk? An empirical examination of risk management
practices in the gold mining industry, The Journal of Finance, 51, 1097-1137.




                                            37

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:11
posted:1/31/2011
language:English
pages:37