The relationship between firm-specific exchange rate exposure and
firms' use of foreign currency derivatives
Bachelor Thesis by Casper Deitch
Faculty of Economics & Business
Economics and Finance
Supervised by Ms. L. Zhao MSc
Table of Contents
1. Abstract
1
2. Introduction
2
3. Literature review
3
3.1 Theoretical Review
4
3.2 Empirical Findings
9
4. Data
15
5. Methods
16
6. Findings
19
7. Conclusion
21
8. References
24
9. Appendix
26
1. Abstract
This study aims to investigate whether firm-specific exchange rate exposure is a determinant of firms' use of foreign currency derivatives (FCDs), for a sample of 42 large European non-financial firms for ten years (2004 - 2013) by regressing FCD use on firm-specific exchange rate exposure, as determined within an augmented capital asset pricing model (CAPM). No significant relationship is found between firm-specific exchange rate exposure and FCD use, suggesting that, contrary to previous findings, firm-specific exchange rate exposure is not a determinant of FCD use.
2. Introduction
Firms allocate managerial and financial resources to the hedging of exposure to various types of risk. This is because exposure to risk causes fluctuations in cash flows, and these fluctuations have undesirable effects on firm value, such as financial distress, additional tax costs and information asymmetry. Financial
derivatives can be used as hedging instruments. However, alternative motives which are not necessarily related to maximizing firm value, such as speculation, also exist. Therefore, it is important for shareholders to know for which purposes these derivatives are being used (Hagelin, 2003). Because of globalization, international trade has increased substantially in recent decades and thus exposure to currency risk has become more widespread and more substantial. Foreign currency derivatives (FCDs) can be used as instruments to hegde this exposure. Foregoing literature has already examined the use of FCDs and generally found a positive relationsip between FCD use and currency exposure. However, this research has mainly been done on data from more than a decade ago. There have been significant disruptions in financial markets and currency markts in recent years. The financial crisis of 2008, for example, highlighted the destructive potential of derivatives, when overvalued or used in abundance, and it is likely that this has affected firms' hedging behavior. This paper will investigate whether firm-specific exchange rate exposure is a significant determinant of the use of foreign currency derivatives (FCDs) for a sample of 42 large firms in the Euro Area, for the years 2004 to 2013. This research will add to the existing knowledge by using more recent data to provide up to date findings in this research area. This study will also contribute in terms of methodology: previous research has generally employed proxy variables to measure exchange rate exposure, and has used a binary variable for FCD use , while this study computes exact exchange rate exposure within an augmented capital asset pricing model (CAPM) and it employs a continuous variable for FCD use, so that additional insights can be gained on the extent to which firms hedge with FCDs. The next section is a review of the literature on this topic. Section 3 is about the data sample and the selection process and section 4 describes the methodology of this study. The findings are presented and evaluated in section 5, and section 6 is provides a summary and conclusion.
3. Literature Review 3.1 Theoretical review
Why do firms engage in hedging activities?
The classic Modigliani and Miller model proposes that with perfect capital markets, the firm's capital structure is irrelevant, and therefore fluctuations in cash flow do not affect firm value and there is no need for hedging activities (Berk & DeMarzo, 2011). This implies that there needs to be a market imperfection, for hedging activities to be justified. Foregoing theoretical literature has identified various types of market imperfections which provide an incentive for hedging. Smith and Stulz (1985, cited in Allayanis & Weston, 2001) identify taxes as such an imperfection: there may be tax benefits associated with maintaining a stable pattern of cash flows. They also find that fluctuations in cash flow may cause financial distress, and the transaction costs associated with this would then reduce firm value. Fluctuations in cash flows may also lead to financial constraints, such that firms may not be able to invest when valuable growth opportunities arise (Lessard, 1990; cited in Geczy et al, 1997). Another imperfection is the costs associated with the
underinvestment problem which may occur when cash flows are uncertain and external financing is costly (Froot, Scharfstein & Stein, 1993, cited in Allayanis & Weston, 2001). Stulz (1984, cited in Allayanis & Weston, 2001) identifies manager's risk aversion as a potential explanation: if firm managers are risk averse they may try to limit cash flow fluctuations by engaging in hedging activities. This is especially the case when they own a large amount of the firm's shares (Smith & Stulz, 1985; cited in Geczy et al, 1997). Note that this imperfection does not relate to firm value, but to managers' personal preferences. However, DeMarzo and Duffie (1995, cited in Allayanis & Weston, 2001) find that hedging benefits increase when hedging is done by managers who have private information on the company's expected profits. Therefore it may be justified for shareholders to let managers engage in hedging activities nonetheless. On the other hand, if managers own options on the firm's stock, they may have an incentive to increase the firm's risk, so that the value of their options increases (Smith & Stulz, 1985; Geczy et al, 1997). In this case they will be less likely to engage in hedging activities. Sharpe (1994; cited in Hagelin, 2003) hypothesized that firms
balance the benefits of leverage with the benefits of a highly trained workforce, so that firms allow reductions in their workforce when leverage increases. The implication of this, is that firms with high employee training costs will be averse to increases in leverage, and therefore will engage in hedging activities to prevent the need for excessive leverage (Hagelin, 2003). Table 1 summarizes the theoretical justifications for hedging against exposure*.
Why do firms use derivatives?
The added value of hedging with derivatives, as opposed to hedging by buying the underlying asset itself, is that derivatives offer leverage, in the sense that a large volume of price exposure can be achieved, with a relatively low initial investment. This is because the hedging party only pays the costs associated with the derivative contract, but obtains exposure to changes in the value of the entire amount of underlying assets covered in the contract (Bodie, Kane and Marcus, 2009). The fact that hedging activities can increase firm value, and that derivatives provide an effective method of doing so, provides a rationale for firms to make use of derivatives. Depending on the type of exposure firms face, there may be other effective alternatives to derivatives. This thesis only investigates derivatives used for foreign currency exposure and the
corresponding alternatives are discussed briefly in the next section.
Foreign Currency Derivatives (FCDs)
The process of globalization has increased international trade substantially in recent decades and thus substantial exposure to currency risk has become commonplace among large firms. This is one reason why it is interesting to investigate the use of FCDs. Another, more practical reason is that investigating the use of derivatives for a specific type of exposure improves the ability to control for variation in the inherent exposure (Hagelin, 2003). The theoretical case for using FCDs to hedge against currency exposure is the same as that for other derivatives: to reduce fluctuations in cash flows, in this case due to fluctuations in the exchange rate. It is important to note that the use of currency derivatives may not be independent of the use of derivatives in general: Geczy et al (1997) argue that economies of scale in hedge management increase the usage of FCDs, because the fixed costs associated with setting up a hedging policy may deter
firms, who make less or no use of derivatives, from investing in FCDs, even when they face substantial currency exposure.
Potential substitutes for FCDs
The literature identifies various potential substitutes for FCDs. Geczy et al (1997) find that foreign
denominated debt can be used as a natural hedge against foreign currency exposure and the research done by Elliott et al (2001) confirms that this is indeed a substitute for the use of currency derivatives. Clarke and Judge (2009) show that FCDs and foreign debt are used as substitutes in the short term, as both methods are deployed by firms to hedge against currency exposure, but as complements in the long term, as foreign debt contracts typically have a longer term than FCDs and therefore, may cause unwanted exposure in the long term, so that supplemental short term contracts, like FCDs, would be required to restore balance. Clarke and Judge (2009) also find that firms with relatively high levels of leverage make less use of foreign debt hedging than firms with low leverage, and that the former prefer to make use of currency swaps, which are a long term FCDs and can therefore serve as alternative to foreign debt hedging. This suggests that high leverage firms are constrained in the use of foreign debt hedging. Furthermore they find that strategies which use only FCDs and strategies which use foreign debt accompanied by FCDs contribute significantly more to firm value than pure foreign debt hedging strategies. They suggest that this is due to certain advantages associated with FCD use, such as low costs, flexibility and the fact that there is no need to access foreign capital markets (Clarke and Judge, 2009). Pantzalis et al (2001) find that firms also make use of operational hedging, when faced with significant currency exposure. Operational hedging is the practice of managing the impact of exchange rate changes on the firm's competitive position across markets, through operational decisions related to marketing, sourcing, plant location, etc. (Pantzalis et al, 2001). Because operational hedging requires long term desion-making and considerable investment, the same flexibility and cost arguments for the advantage of hedging with FCDs apply. Allayanis et al (2001) find that operational hedging is indeed not an effective substitute for financial hedging, but it can be effective as a complementary strategy. This result is confirmed later in a more comprehensive study by Kim et al (2006). The above mentioned hedging alternatives can be viewed as direct substitutes to FCDs, in the sense that
they can contribute to reducing fluctuations in cash flows. However, alternative solutions to the market imperfections, which make fluctuations undesirable in the first place, could be seen indirectly as substitutes for FCDs. For example, Nance et al (1993; cited in Hagelin, 2003) argued that costs of financial distress can not only be mitigated by hedging activities, but also by holding more liquid assets and reducing dividend yield. The issuance of convertible debt and preferred stock are also instruments which could be used to control for financial distress costs (Nance et al, 1993; cited in Geczy et al, 1997). These strategies are, therefore, potential substitutes for hedging instruments in general, and thus also for FCDs. However, these indirect substitutes have their own set of potentially unwanted implications for firm value. For example reducing dividend yield may be an unpopular measure among shareholders and holding more liquid assets could be costly. Furthermore, none of them offer a comprehensive solution to all imperfections, which could eliminate the aversity to fluctuations in cash flow. This suggests that, given the available alternatives, there is still a rationale for firms facing currency exposure, to hedge with FCDs. Table 2 provides a summary of all the above mentioned substitutes for FCDs.
Classification of FCDs and foreign currency exposure
FCD contracts come in various forms. There are currency futures, currency forwards, currency options and currency swaps. Generally the first three types are short term and swaps are long term contracts (Geczy et al, 1997). Foreign currency exposure comes in two forms: translation exposure and transaction exposure. Transaction exposure is the risk arising from potential changes in future cash flows as a result of
unanticipated changes in the exchange rate. Translation exposure is the risk arising from potential changes in the value of the balance sheets of foreign subsidiaries as a result of unanticipated changes in the
exchange rate (Hagelin, 2003). There is a general consensus in the literature that the theoretical arguments for hedging only apply to transaction exposure and not to translation exposure, because translation gains and losses are mostly unrealized and therefore don't affect firms cash flows significantly. Also, translation gains and losses are poor estimators of firm value, so that hedging against them tends to be inefficient in reducing exposure (Eiteman et al, 1995; Sercu and Uppal, 1995; Butler, 1999; all cited in Hagelin, 2003). Transaction exposure is diverse in its nature. Geczy et al (1997) argue that firms who have exposure
resulting from foreign operations and import competition, will primarily make use of short term contracts, such as currency fowards, currency futures and currency options, because this type of exposure is
associated with frequent and uncertain transactions, so that firms will want to be able to adjust and
manage their position actively. However firms who have exposure resulting from foreign debt contracts, will know their exposure at the inception of the contract, so that they will tend to hedge with currency swaps, which generally have a longer term (Geczy et al, 1997). Kim et al (2006) distinguish between transaction exposure and economic exposure. This is similar to the distinction made by Hagelin (2003): transaction exposure represents short term exposure to fluctuations, due to exports, for example. Economic exposure represents structural, long term exposure, due to foreign operations, for example. Firms will generally offset their economic exposure by operational hedging (Kim et al, 2006) or other long term alternatives, such as foreign debt (Eliott et al, 2001), and they will generally manage their transaction expose with FCDs (Kim et al, 2006). Table 3 provides an oversight of the types of currency exposure identified in the literature and the corresponding types of FCDs.
Speculation
All justifications for FCD use mentioned so far are consistent with the theory of hedging. However, there is another reason why firms may want to use derivatives, which is not related to hedging: speculation. Speculation is done to profit from price changes which are not anticipated by the market. However speculation is a relative concept: when fund managers incorporate their market view in their hedging programs this is already a mild form of speculation, called “selective hedging” by Stulz (1996; cited in Adam & Fernando, 2006). Consistent excess volatility in firms' hedge ratios is a sign of this (Adam & Fernando, 2006). According to Fama's (1970) classic doctrine of efficient financial markets, all information available in the market is reflected in current prices, and therefore speculation is value-neutral. In the absence of inside information on firm value, speculation adds risk, since investors are not able to predict how prices will move. So if we incorporate the value effects of the market imperfections mentioned above, which justify hedging, speculation would even be value-decreasing, because it would increase volatility and the aforementioned costs associated with this (Aabo, 2007). However, the literature has identified various
scenario's in which speculation could be value-increasing. As Stulz (1996; cited in Aabo, 2007) points out, it can be value-increasing for shareholders in times of financial distress, to speculate at the expense of debtholders and it can be value-increasing for a firm which is very specialized in a specific market and therefore has superior knowledge on price movements. A study by Adam and Fernando (2006) provides empirical evidence of this for firms in the gold mining industry. Aabo (2007), however, provides two arguments on why this would not be the case for non-financial firms hedging in currency markets. Firstly, they are minor players in the foreign exchange market and therefore it is unlikely they would have
specialized knowledge about future price movements. Secondly, if fund managers do possess such superior knowledge, it is unlikely they would work for non-financial firms, as they could get higher compensation for such skills in financial firms (Aabo, 2007). There are also theories in the literature related to the systematic mispricing of assets, due to inefficencies, such as heterogeneity of information in financial markets, which leads to a bias in prices. For example, Miller (1977; cited in Mei et al, 2005) explains that if market opinions are divergent, stock prices will be upwardly biased because optimists will purchase stocks while pessimists will stay out of the market completely, rather than actively short selling the stock. Such biases in asset prices could enable speculating fund managers to beat the market and increase firm value. Thus there is not yet a general consensus in the literature on whether speculation is theoretically decreasing,
value-increasing or value-neutral. If firms engage in speculative activities, this would be an additional reason for firms to use FCDs, which would be unrelated to hedging motives, and therefore also unrelated to foreign currency exposure. It is important to note that speculation itself, as mentioned above, increases volatility and therefore, it may also increase use of FCDs for hedging purposes. In fact, if firms use FCDs to speculate, it is probable that this will increase foreign currency exposure. However the extent to which speculation causes currency exposure, relative to the extent to which foreign operations cause currency exposure, may be limited for most non-financial, large, multinational firms.
3.2 Empirical findings
Hedging motives and the use of FCDs
Nance et al (1993; cited in Allayanis & Weston, 2001) have found, using survey data, that Fortune 500 firms which use derivatives in general face more convex tax functions, indicating that there are indeed tax benefits associated with the use of derivatives. Nance et al (1993; cited in Hagelin, 2003) hypthesized that the probability of financial distress increases with leverage. Hagelin (2003) argues that growth opportunities can be used as a proxy for the underinvestment problem, as firms with more valuable growth opportunities and higher leverage are more likely to suffer from it. Thus, the research of Hagelin (2003) uses leverage, measured by the book value of debt divided by the book value of equity, as a proxy for financial distress costs and it uses growth opportunities, measured by the ratio of the market to the book value of total assets and firm size, as a proxy for the occurrence of the underinvestment problem. His research confirms the positive relationship between the use of FCDs and costs of financial distress as well as the underinvestment problem. The research of Geczy et al (1997) confirms this positive relationship between FDC use and future growth opportunities, but it additionally uses the ratio of research and development costs to sales and the ratio of capital expenditures on property, plants and equipment to firm size, as proxies for growth
opportunities. They also find that the use of FCDs is more likely among firms with tighter financial constraints. Tufano (1996, cited in Allayannis & Weston, 2001) found evidence amongst US gold mining firms on the relative amount of options and stocks held by managers, consistent with the managerial risk aversion hypothesis, indicating this is indeed an incentive for managers to use commodity derivatives. Empirical research om the relationship between managerial stock or option holdings and the use of
derivatives is mixed. Berkman and Bradbury (1996; cited in Nguyen & Faff, 2002) find systematic consistency between managerial holdings of stock and options and hedging patterns for US gold mining firms. However, Haushalter (2000; cited in Nguyen & Faff, 2002) finds no evidence of a relationship between managerial stock holdings and FCD use, and mixed evidence with regard to managerial option holding. The research of Nguyen and Faff (2002) produce a significant relationship neither between managerial stock holding and
derivatives use, nor between managerial option holding and derivatives use. The fact that managers may have private information on the company's expected profits implies that shareholders may encourage them to engage in hedging activities. Therefore greater information asymmetry should increase the use of derivatives (Demarzo & Duffie, 1991; cited in Geczy et al, 1997). Geczy et al (1997) have researched this relationship for FCD use specifically. They used the firm's level of institutional ownership and the amount of analysts following the firm, as proxies for information availability: the opposite of information asymmetry. Their research produces a negative relationship between these variables and FCD use, which confirms the positive relationship between information asymmetry and the use of FCDs. However, Graham and Rogers (2000; cited in Hagelin, 2003) found a positive relationship between institutional ownership and derivatives use in general. The findings of Hagelin (2003) confirm this positive relationship for currency derivatives. This would imply that firms with less information asymmetry are more likely to hedge with (currency)
derivatives. Thus, the evidence on this effect is also mixed. Finally, Hagelin (2003) proxies for employee training costs with the level of human capital investments. However he finds no significant relationship between this variable and FCD use. There is no further significant research on this effect.
Economies of scale and the use of FCDS
Geczy et al (1997) find that the use of FCDs is positively related to firm size and the use of other hedging instruments, indicating that FCD usage is positively related to economies of scale.This result, especially with regards to firm size, is generally confirmed in the literature (Nance et al, 1993; Sinkey & Carter 1994;
Cummins et al, 1997; all cited in Gonzales et al, 2007). Gonzales et al (2007) find that scale economies are one of the main determinants of FCD use.
Substitutes and the use of FCDs
Kim et al (2006; cited in Clark & Judge, 2009) find that though the use of FCDs increases firm value,
operational hedging can increase firm value up to five times as much. However, Allayannis et al (2001; cited in Clark & Judge, 2009) find that operational hedging only increases firm value when combined with FCDs, suggesting that operational hedging and FCD use are complements. Geczy et al (1997) use R&D
expenditures and short-term liquidity to proxy for foreign operations, because firms with high R&D
expenditure are usually forced to seek revenues abroad when domestic financing becomes too costly, even though this typically causes a mismatch between costs and revenues. This mismatch provides an incentive for FCD hedging. Short-term liquidity is an indication of the availability of funds and therefore should be inversely related to mismatching foreign operations. They find that for firms with foreign operations and foreign denominated debt these proxy variables are not significantly related to FCD use, but that for firms with foreign operations and no foreign debt they are significant determinants of FCD use. This suggests that foreign denominated debt is indeed a natural substitute to FCD use. Three studies done by Nance et al (1993; cited in Hagelin, 2003) found a negative relationship between liquidity and hedging with derivatives. The research done by Gonzales et al (2007) confirms the siginificance of this effect. However, Hagelin (2003) did not find such a relationship for the use of FCDs. None of the studies done by Nance et al (1993; cited in Hagelin, 2003) found the hypothesized positive relationship between dividend yield and derivatives use and neither did Hagelin for FCD use (2003). The research of Geczy et al (1997) also fails to confirm the dividend yield hypothesis for FCDs. This research also fails to confirm the hypothesized relationships of convertible debt and preferred stock issuance with FCD use. Table 4 provides an overview of the empirical evidence on all determinants of FCD use for hedging purposes, other than currency exposure.
Foreign currency exposure and the use of FCDs
Hagelin (2003) collects survey data, from a sample of Swedish firms in the year 1996, on FCD use and on
which type of currency exposure those firms are hedging against. Hagelin (2003) primarily uses FCD use as a binary variable, to distinguish between users and non-users. To measure exchange rate exposure he uses two proxy variables: foreign revenues and foreign equity. A logit regression is performed to determine the relationship between these proxies and the likelihood for a firm to use FCDs. Furthermore, the survey data is used to distinguish between firms which hedge against translational exposure, transactional exposure to committed transactions and transactional exposure to anticipated transactions. For each type a separate logit regression is performed. The result of this research is that the likelihood of FCD use does indeed increase with foreign currency exposure, but only as measured by foreign equity. Furthermore, this
significant relationship holds for transactional exposure, but not for translational exposure. Hagelin (2003) finds from univariate tests that firms which hedge translational exposure have a more diverse hedging approach than those which hedge transactional exposure. To obtain additional insights in the hedging behavior of this group Hagelin (2003) employs Cragg's (1971) model which combines a probit analysis with a truncated regression. The probit analysis incorporates the decision to hedge and the truncated regression is the regression equation for non zero outcomes. The outcome of the probit analysis is consistent with the initial logit regression. From the truncated regression it is found that the level of FCD use is significantly positively related to currency exposure, even for firms which hedge against translational exposure.
Geczy et al (1997) do a cross-sectional study on US non financial firms for the year 1990 to examine the
determinants of corporate derivatives use. This research also takes FCD use as a binary variable and also runs a logit regression on proxy variables for currency exposure. In this case a greater amount proxy variables are used: the ratio of pre-tax foreign income to total sales, the ratio of foreign sales to total sales, the ratio of foreign assets to total assets, the dollar-equivalent amount of long- and short-term foreign denominated debt, a binary variable which indicates whether or not a firm has foreign denominated debt and the percentage of imports in a firm's industry, relative to total industry output. They find that the use of FCDs is indeed positively related to foreign exchange exposure. Their research also confirms that firms who have exposure resulting from foreign operations and import competition, are more likely to use currency forwards, futures or options and that firms who have exposure resulting from foreign debt contracts are more likely to use currency swaps.
Gonzales et al (2007) conduct their research on a sample of 49 non-financial Spanish firms in 2003. Their
research uses a limited variable for FCD use with a corner solution. They choose to run a tobit regression, because it's restricted structure is appropriate for such a variable. However, according to Gonzales et al (2007) the decision whether to hedge is not affected by foreign exchange exposure. Only the extent to which they use FCDs is affected by this. Therefore they do not include a variable for exchange rate exposure in their tobit regression. In stead, they conduct the aforementioned analysis from Cragg (1971), in the same
way as Hagelin (2003). For the corresponding probit analysis, concerning the decision to hedge, they do not include an exchange rate exposure variable. They include the variable foreign sales in the truncated
regression, concerning the hedging volume, as a proxy for exchange rate exposure. They find that this variable is the greatest determinant of the hedging volume, suggesting a strong positive relationship between foreign currency exposure and FCD use.
Speculation and the use of FCDs
The findings of Geczy et al (1997) on determinants of FCD use, are consistent with optimal hedging behavior theories and inconsistent with speculative motives. The variables they use as proxies for speculative motives are, firstly, firm size and use of other derivatives, because these proxy for economies of scale in transaction costs, which could allow profitable arbitrage opportunities. Secondly, they use variables for financial distress, because managers who hold options on the firm's value or who hold firm shares and view them as options on firm value, may want to speculate to increase volatility when the firm is near or in financial distress. Finally, it is possible that firms with low output have a signaling incentive to speculate: if they can increase volatility by speculation, their output may mimic that of high output firms. However this last theory relies on the assumption that speculation is unobservable. Firms with low output are generally close to or in financial distress, so that they are likely to be monitored by outside debt holders, which makes this assumption unrealistic. Even though Geczy et al (1997) find that economies of scale are positively related to FCD use, this result is also consistent with optimal hedging theories and therefore does not provide
evidence for speculation. They also find that financial distress is not significantly related to FCD use. Furthermore, they find that the firms in their sample without ex ante exchange rate exposure, do not use FCDs. This evidence suggests that firms, on average, do not use FCDs for speculation. Gonzales et al (2007) find that the FCD use of firms in their sample matches the levels of currency risk which those firms face, which is further evidence that firms, generally, do not use FCDs for speculation. However, a survey by Bodnary and Gebhardt (1998, cited in Gatopoulos and Louberge, 2013) reveals that a significant percentage of firms uses derivatives for speculative activities, as a result of their 'market view'. Dolde (1993; cited in Aabo, 2007) and Glaum (2002; cited in Aabo, 2007) also find empirical evidence that firms use derivatives
speculatively for US and German firms, respectively. Allayannis et al (2003; cited in Allayannis et al, 2012) have also found evidence among East Asian firms of speculation: their derivatives use was affected by various macro-factors in a manner consistent with speculation, rather than hedging. Recent empirical evidence on speculation, it's merits and the frequency of its occurrence in firms is scarce. This may be due to the recent financial crisis, which has made speculation a controversial topic. It is a fact that all firms in the sample of this thesis paper report that they do not make use of derivatives for speculative purposes.
However, this may not necessarily reflect reality.
Generally, foregoing research agrees that the use of FCDs is a result of firm's foreign exchange risk exposure and that they are used for hedging against this risk, rather than for speculating. However, these conclusions have predominantly been reached from researching proxy variables. Also the research has been done mainly on cross-sectional data from one year, rather than on time-series data. Finally, most research has been done on firms in the 1990s. Gonzales et al (2007) is an exception to this, though this research was done only on Spanish firms in the year 2003. This thesis will contribute to the literature in three ways. Firstly it applies a more exact method of determining exchange rate exposure, through an augmented capital asset pricing model (CAPM), so that findings do not rely on assumptions regarding the represenativeness of a certain proxy variable. Secondly, panel data has been collected so that research could be conducted on multiple firms, from multiple countries (within Europe) and their activity over multiple years. Finally, the data is recent, from 2004 to 2013, and therefore includes observations from before, during and after the 2008 financial crisis, which has made derivatives such a controversial topic.
4. Data
Data has been collected on a sample of 42 large European firms, listed on major stock exchanges. They were selected from a list of the largest European firms, as measured by revenue. The selection procedure started from the top of the list, with the largest firm. Financial companies were excluded from the list, because due to the specific nature of their industry, they hold financial assets, such as FCDs, for trading purposes. Including these in the sample would produce biased results, by overstating the amount of FCDs firms generally use. The next step was to read through the financial statements of the annual reports of each firm for each year from 2004 to 2013, to obtain the corresponding data on FCD use. FRS 13 mandates the
disclosure of the fair values of derivative contracts, as well as the direction of their position. However, apparently not all firms have adopted this practise. This is why not all non financial firms of the list could be included in the sample. Furthermore, some firms only included the notional amounts of their FCD holdings. This data was collected initially, but has not been included in the sample, because these values by
themselves do not provide a clear picture of the extent nor the direction of exposure (or coverage of exposure). Finally, a couple of firms were excluded for practical reasons, for example, annual reports not available in English, not listed on a stock exchange, etc. Out of the top 78 companies on the list, 42 companies fit the criteria and provide the fair values of their FCD holdings in their annual reports. Some firms have not started including FCD data in their annual reports until after 2004, and some firms have not existed until after 2004, so for these firms there is less than 10 years of data in the sample. Apart from these missing values, the eventual sample contains 2608 days of observations for 42 companies. In order to estimate the exchange rate exposure in the augmented CAPM, daily stock price data for each of these companies was sourced from Datastream, along with the daily values of the value-weighted EMU market index over the same period. The daily values of the three exchange rates used in the model were sourced from Compustat, because Datastream was not accessible to the author at this point. The three currency rates employed as currency factors are the Euro to US Dollar rate, the Euro to British Pound rate and the Euro to Japanese Yen rate. These rates were chosen because the US, the UK and Japan are the biggest
trading partners of EU companies (Muller and Verschoor, 2006).
5. Methods
The purpose of this study is to investigate whether firm-specific exchange rate exposure is a determinant of FCD use. Previous research has generally found that this is the indeed the case, because patterns of FCD usage are consistent with optimal hedging behaviour, rather than with optimal speculative motives. It has been shown that currency rate exposure is a significant determinant of FCD use. Hence the null hypothesis and alternative hypothesis of this study are as follows:
Ho: Firm-specific exchange rate exposure is not a significant determinant of FCD use.
H₁: Firm-specific exchange rate exposure is a significant determinant of FCD use.
This hypothesis is based on the theory of optimal hedging behavior, which proposes that firms who face risk from exposure to currency fluctuations, will offset their currency position with FCDs.
Annual exposure levels are derived for each firm over a period of ten years. This is done by regression analysis with an augmented CAPM equation, following the methodology of Muller and Verschoor (2006). The augmented CAPM equation has the following form:
Ri(t) = αi + βi*Rm(t) + γi*X(t) + εi(t)
where Ri(t) is the total return* of firm i in period t, βi is the sensitivity of firm i's total return to market fluctuations, Rm(t) is the total return* of the european stock market, γi is firm i's firm-specific exposure to exchange rate changes, independent of the effects these may have on market return, X(t) is the rate of return* on three of the most actively traded currencies vis-a-vis the euro
the findings of this thesis can be compared to their findings.
in period t and εi(t) denotes the value of the error term in period t. Note that γ is firm-specific, and therefore it excludes the effect of exchange rate fluctuations which affect the market as a whole, because this effect is captured in β, the coefficient of the market factor. The stock-, market index- and exchange rate returns are computed from daily stock-, market index- and exchange rate values by the following formula:
Ri(t) = (Pi(t) -Pi(t-1))/Pi(t)
Where Ri(t) is the daily return of i on day t and Pi(t) is the value of i on day t. The annual coefficients of each firm are estimated by running an OLS regression of firm i's daily returns for one year on the daily market returns factor and currency returns factor, for that same year. In order to explore all possibilities with regards to the currency effect, first the regressions are carried out for each currency separately to find the exposures of firms to each currency. Then the model is expanded to contain all three currency returns factors, so that three exposure figures are produced per regression. This is done, because a firm may be exposed to currency risk from more than one currency, and therefore a regression containing only one currency factor among its explanatory variables may suffer from omitted variable bias. Finally, a currency returns factor is generated which is the sum of the three currency returns. This is because, if a company is significantly exposed to more than one of these currencies, the effects of changes in these rates may cancel out (or add up). The joint significance of the regression model is determined through an F-test and the significance of each γ is determined through a t-test. The t-test has the following form:
t = γi(T) / SE(γi(T))
Where, γi(T) is the currency exposure of firm i in year T and SE(γi(T)) is the standard error of the currency exposure of firm i in year T. Note that being exposed to currency fluctuations is not the same as being affected by currency fluctuations: a firm which is exposed to currency fluctuations will not be affected untill
the currency actually fluctuates. In other words, currency exposure, as measured in this study is only the sensitivity of the firm's market value to a potential change in the value of a relevant currency rate. This sensitivity depends on operational and financial variables, such as foreign sales, foreign debt, etc. The actual effect of a currency rate change on the market value of a firm is determined by multiplying the change in exchange rate by the firm's corresponding exposure to that rate. The fact that firm-specific exposure is related to such operational and financial variables, which tend to change significantly only in the long term, justifies the choice to derive annual exposure figures: it is reasonable to assume that the extent to which firms are involved in foreign markets is relatively constant throughout the year.
In the second part of this study, the annual γ coefficients are regressed on the annual FCD holdings of each company, to determine whether there is a relationship between these variables. The significance of this relationship will be investigated again by an F-test for the whole regression and a T-test on each coefficient. Prior research generally has employed a binary variable to represent FCD usage (see literature review). However, in the sample which was collected for this study nearly all firms reported FCD holdings. Thus, using a binary variable would result in non-users being under represented in the sample. However, there is added value in being able to use a continuous variable and incorporate the exact amounts of FCD holdings for each firm, as this can provide information on the extent to which firms use FCDs to hedge against exposure.
6. Findings
Tables 5, 6 and 7 summarize the findings of the first regression for the three separate currency returns factors and table 8 summarizes the findings for the combined total currency returns factor. The above mentioned tables show the annual γ of each firm. Table 9 shows the results, which the regression yields, when the whole ten year period is utilized to produce one γ per company. The figures in bold are significant at at least a 10% significance level. Missing data is represented by a dot. Furthermore, for each significant γ coefficient, the corresponding F-statistic indicates that the model as a whole is significant at 1%. Because of the relatively small amount of significant γ coefficients found, coefficients with a p-value between 10% and 15% percent are also included in the tables.
The percentages of significant γ coefficients found for each currency factor are 6.0%, 6.2%, 16.9% and 21.0% for, respectively, the US Dollar, the British Pound, the Japanese Yen and the total currency factor. This would suggest that the majority of firms in the sample are not exposed to currency fluctuations. This is not a realistic result, as the firms in the sample are all large, multinational organizations. Muller and Verschoor (2006), find similar percentages of currency exposure in their study. There are various reasons why such a small amount of significant exposures is found. Muller and Verschoor (2006) mention the intervaling effect among investors: in this model the currency exposure is determined using the change in market value of the firm as measured by stock price returns. This value is a reflection of investors' beliefs and they may make errors in their predictions of the long term effects of currency rate changes: an expected effect of such a change is immediately incorporated into the stock price by investors' valuation, while the effect of this change on actual firm value takes longer to occur and therefore it is difficult for investors to accurately predict it. For further research a possible solution to this could be to take returns over a longer time period than daily. Muller and Verschoor (2006) have found greater percentages of significant exposure for returns over longer periods of time. However, the sample size would need to be increased: for example, changing
from daily returns to monthly returns would reduce the sample size by a factor of 30. Another reason for the low percentages of exposure could be the fact that some firms, for some periods, are naturally or operationally hedged. This would cancel out their effective exposure. This would also imply that these firms, would make less use of currency derivatives, because there would be less need for this. For further
research, the companies for which this is the case could be identified by using variables which proxy for operational and natural hedging (see literature review). Finally, it is possible that investors expect
companies to actively hedge against currency fluctuations to such an extent that they immediately adjust their hedging portfolio to maintain a zero-exposure position throughout the year. If this is the case, stock prices may not change at all as a result of exchange rate changes. However, this raises the question why significant exposure is found for various companies for various years. It could be the case, that in these years, investors temporarily lose faith in these firms' hedging capabilities. In support of this argument, significantly more exposure figures were found during 2008 and 2009, the peak years of the financial crisis, than in other years. The financial crisis of 2008 is commonly associated with information asymmetry and low confidence in financial markets. Such information problems may affect investor's confidence in firms' ability to hedge and therefore, they may be reflected in investors' estimation of exchange rate effects on firm value. The effect of information problems on investors' estimation of firms' hedging capabilities is also a topic for further research which could help to improve the current model for determining exchange rate exposure.
Because of the limited amount of significant (at 10%) coefficients, all coefficients with a p-value equal to or under 15% were utilized in the second part of the study, which is the regression of FCD use on exposure. This resulted in sample sizes of 33, 34, 84 and 119 for the exposures to, respectively, US Dollar, British Pound, Japanese Yen and the total currency factor (TOTCUR). Table 9 shows some basic summary statistics of firms' FCD use and currency exposure. To get a meaningful average the absolute values were used. Also, some outliers had to be removed from the exposure data, to get representative summary statistics. As nearly all exposure figures were between 0 and 1, all figures above 1 were removed for this table.
Table 10 shows the correlations between firms' annual FCD use (FCD) and their corresponding annual exposure to each currency factor and table 11 summarizes the regression coefficients of the regression of FCD use on currency exposure. The correlations between FCD and each currency factor are very low and the regressions for each currency factor are insignificant as measured by the p-value of the corresponding F-tests. Also the corresponding coefficients of the currency factors are insignificant as measured by the p-values of the corresponding t-tests. This would suggest, contrary to all earlier findings that exchange rate exposure is unrelated to FCD use. This result would imply that firms use FCDs for speculation, rather than hedging. However, this seems unlikely. Firstly, as mentioned in the literature review, non-financial firms are not specialized in foreign exchange markets and therefore are unlikely to have the kind of information advantage which would make speculation profitable. Secondly, even if firms were to speculate with FCDs, they would still have hedging incentives. So it would be likely that they would engage in 'selective hedging' (see literature review), rather than outright speculation. However, in the case of 'selective hedging' a relationship should still be found between currency exposure and FCD use. Because such a relationship is not found, and because outright speculation on foreign exchange markets is unlikely for the firms in this sample, other explanations are required for the outcomes of this regression. One explanation could be the fact that the sample size, which was significantly reduced after the first regression, due to insignificant coefficients, is too small to find a significant effect. This problem could be mitigated for future research, by improving the currency exposure model in the ways mentioned above, or collecting more data. However, the latter would be time-consuming, so that either more time or more researchers would be required for this. Another explanation could be omitted variables: if FCD use is viewed as a reflection of general hedging behaviour, it may be influenced by a variety of other factors, such as the ones mentioned in table 1 of the literature review. If firms make significant use of direct and indirect substitutes, such as the ones mentioned in table 2 of the literature review, this could also bias the outcomes of the regression. Note that widespread use of substitutes to FCD use is theoretically unlikely due to higher costs and lower flexibility (see literature review). Omitted variable bias results in an overstatement of the standard deviation of regression
coefficients, leading to an understatement of the corresponding t-statistic. This could explain why none of the coefficients were significant. In future research this could be prevented by adding the above mentioned
variables to the regression equation. Finally, another explanation could be the fact that annual FCD holdings have been taken from firm's financial statements and thus are point-in-time data, while annual exposure has been determined from daily returns throughout the year. This was done on the assumption that exposure is relatively constant throughout the year (see methods), and therefore FCD holdings, if used to cover this exposure should be relatively constant as well. If this assumption is not realistic, there could be a mismatch between the data of the two variables: the exposure figures would represent an average,
determined by inputs which vary day by day, while the FCD variable would represent FCD use, only for that point in time when the financial statements are drafted. This is an issue which will be hard to resolve in future research, as companies do not disclose their daily FCD holdings.
The use of a total currency factor could be seen as a contribution the literature on this subject. It is something which has not been done yet, and it produced more significant exposure coefficients than the separate currencies did in the augmented CAPM.
7. Conclusion
The goal of this study was to investigate whether firm-specific exchange rate exposure is a significant determinant of FCD use. Accordingly the null hypothesis and alternative hypothesis of this study are as follows:
Ho: Firm-specific exchange rate exposure is not a significant determinant of FCD use.
H₁: Firm-specific exchange rate exposure is a significant determinant of FCD use.
No significant results were found to reject the null hypothesis, suggesting that firm-specific exchange rate exposure is not a significant determinant of FCD use. Various reasons have been suggested for this unexpected outcome: speculative use of derivatives, limited sample size due to the limited amount of significant currency exposure data produced by the first part of the study, omission of variables related to hedging behaviour, omission of variables related to substitutes and a potential mismatch of data between variables. Future research could focus on improving the model for exchange rate exposure, for example by incorporating the effects of information asymmetry between investors and firms and investor confidence in the model. Another topic for further research could be how firms' use of FCDs varies throughout the year. The fact that firms only disclose their holdings of FCDs for the day that their financial statements are drafted, is a major limitation to this area of research.
8. References
Aabo, T. (2007). The impact of individual-owners on currency speculation: the case of Danish non-financial firms. International Journal of Managerial Finance, 3, (1), 92-107.
Adam, T. R. & Fernando, C. S. (2006). Hedging, speculation and shareholder value. Journal of Financial
Economics, 81, 283-309.
Allayannis, G., Ihrig, J. & Weston, J. P. (2001). Exchange-Rate Hedging: Financial versus Operational Strategies. The American Economic Review, 91, (2), 391-395.
Allayannis, G., Lel, U. & Miller, D. P. (2012). The use of foreign currency derivatives, corporate governance, and firm value around the world. Journal of International Economics, 87, 65-79.
Allayannis, G. & Weston, J. P. (2001). The Use of Foreign Currency Derivatives and Firm Market Value. The
Review of Financial Studies, 14, 243-276.
Berk, J. & DeMarzo, P. (2011). Corporate Finance: Global Edition. New York: Pearson.
Clark, E. & Judge, A. (2009). Foreign Currency Derivatives versus Foreign Currency Debt and the Hedging Premium. European Financial Management, 15, (3), 606-642.
Elliott, W. B., Huffman, S. P. & Makar, S. D. (2003). Foreign-denominated debt and foreign currency derivatives: complements or substitutes in hedging foreign currency risk? Journal of Multinational
Financial Management, 13, 123-139.
Fama, E. F. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work. Journal of Finance, 25, (2), 383-417.
Gatopoulos, G. & Loubergé, H. (2013). Comined use of foreign debt and currency derivatives under the threat of currency crisis: The case of Latin American firms. Journal of International Money and
Finance, 35, 54-75.
Geczy, C., Minton, B. A. & Schrand, C. (1997). Why firms use currency derivatives. The Journal of Finance, 52, (4), 1323-1354.
Gonzales, L. O., et al. (2007). Why Spanish Firms Hedge with Derivatives: An Examination of Transaction Exposure. SSRN Working Papers, No. 1003358.
translation exposure. Applied Financial Economics, 13, 55-69.
Kim, Y. S., Mathur, I. & Nam, J. (2006). Is operational hedging a substitute for or a complement to financial hedging? Journal of Corporate Finance, 12, 834-853.
Mei, J., Scheinkman, J. & Xiong, W. (2005). Speculative trading and stock prices: evidence for Chinese A-B share premia. NBER Working Papers, No. 11362.
Muller, A. & Verschoor, W. F. C. (2006). European Foreign Exchange Risk Exposure. European Financial
Management, 12, (2), 195-220.
Nguyen, H. & Faff, R. (2002). On The Determinants of Derivative Usage by Australian Companies. Australian
Journal of Management, 27, (1), 1-24.
Pantzalis, C., Simkins, B. T. & Laux, P. A. (2001) Operational Hedges and the Foreign Exchange Exposure of U.S. Multinational Corporations. Journal of International Business Studies, 32, (4), 793-812.
9. Appendix
Table 1: Theoretical justifications for hedging
1 Tax costs Smith and Stulz, 1985
2 Costs of financial distress Smith and Stulz, 1985
3 Financial constraints Lessard, 1990
4 Underinvestment problem Froot, Scharfstein & Stein, 1993
5 Managers' risk aversion Smith & Stulz, 1985
6 Managerial stock holdings Smith & Stulz, 1985
7 Private information among managers DeMarzo & Duffie, 1995
8 High employee training costs Hagelin, 2003
Table 2: Substitutes for FCD use Direct substitutes
1 Operational hedging Kim et al, 2006
2 Foreign denominated debt Geczy et al, 1997
Indirect substitutes
3 Holding more liquid assets Nance et al, 1993
4 Reducing dividend yield Nance et al, 1993
5 Issuance of convertible debt Nance et al, 1993
6 Issuance of preferred stock Nance et al, 1993
Table 3: Currency exposure types and FCD types
Transaction exposure
(Hagelin, 2003) resulting from foreign operations or import competition Currency forwards, futures or options Geczy et al, 1997 resulting from foreign
denominated debt Currency swaps Geczy et al, 1997
Transaction exposure
(Kim et al, 2006) All types of FCDs Kim et al, 2006
Translation exposure None Hagelin, 2003
strategies
Table 4: Empirical support for findings on FCD use for hedging
Theoretically positively related to FCD use Relationship empirically confirmed by
Tax costs Nance et al (1993)
Costs of financial distress Hegelin (2003); Geczy et al (1997)
Financial constraints Geczy et al (1997)
Underinvestment problem Hagelin (2003); Geczy et al (1997)
Manager's risk aversion Tufano (1996)
Managerial stock/option holdings Mixed evidence
Private information among managers Mixed evidence
High employee training costs No evidence
Economies of scale Geczy et al (1997); Nance et al (1993); Sinkey & Carter
(1994); Cummins et al (1997); Gonzales et al (2007)
Speculation Mixed evidence
Theoretically negatively related to FCD use Relationship empirically confirmed by
Operational hedging Mixed evidence
Foreign denominated debt Geczy et al (1997)
Holding more liquid assets Mixed evidence
Reducing dividend yield Mixed evidence
Issuance of convertible stock No evidence
Table 5: Firm specific exchange rate exposure to the US dollar Company name 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 HEINEKEN HLDG. -.0112¹ -.0074⁵ TOTAL .0039¹⁵ GAZPROM .0298⁵ E ON -.0066⁵ GDF SUEZ .0073¹⁰ -.0178⁵ FIAT -.0176⁵ -.0122¹⁰ OIL COMPANY LUKOIL -.0162⁵ .0339⁵ ENEL .0043¹³ TESCO -.0196¹ -.0053¹⁰ BASF -.0043¹⁰ -.0073⁵ ARCELORMITTAL -.0072¹³ -.0127⁵ METRO -.0179¹⁰ EDF -.0159⁵ PEUGEOT -.0153¹⁰ -.0087¹³ DEUTSCHE TELEKOM -.0080¹ .5599¹ -.0079¹³ DEUTSCHE POST -.0069¹⁰ THYSSENKRUPP -.0062¹² -.0097¹⁰ RWE -.0063¹² AIRBUS GROUP -.0131⁵ -.0089¹⁴ RIO TINTO -.0127⁵ -.0068¹⁵ BAYER -.0066¹⁰ SSE -.0106¹⁰ VOLVO 'B' -.0068¹² ¹ Significant at 1% ⁵Significant at 5% ¹⁰Significant at 10%
¹¹Would have been significant at 11% level ¹²Would have been significant at 12% level ¹³Would have been significant at 13% level ¹⁴Would have been significant at 14% level ¹⁵Would have been significant at 15% level
Table 6: Firm specific exchange rate exposure to the British Pound Company name 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 HEINEKEN HLDG. -.0114¹
-.0070¹⁰ BP -.3995⁵
TOTAL .0040¹⁴ GAZPROM .0318⁵ E ON -.0071⁵ GDF SUEZ -.0176⁵ FIAT -.0182⁵ -.0124¹⁰ SIEMENS -.2033¹³
OIL COMPANY LUKOIL -.0158⁵ .0362⁵ -.0089¹⁵ ENEL .0049¹³ TESCO -.0199¹ -.0058¹⁰ BASF -.0044¹⁰ -.0078⁵ ARCELORMITTAL -.0140¹ NESTLE 'R' -.2063¹²
METRO -.0166¹⁰ EDF -.0151¹⁰ PEUGEOT -.0155¹⁰ -.0089¹³ DEUTSCHE TELEKOM
.5609¹
-.0089¹⁰ DEUTSCHE POST -88.7654¹
-.0074¹⁰ VODAFONE GROUP -.2925¹⁰
THYSSENKRUPP -.0118¹¹ RWE -.2896¹¹
AIRBUS GROUP -.0150¹ -.0094¹¹ ORANGE -.0070¹⁵ RIO TINTO -.0124⁵ -.0075¹⁴ NOVARTIS 'R' -.2142¹³
VINCI .4477¹
BAYER -.0063¹¹
SSE -.0103¹¹
¹ Significant at 1% ⁵Significant at 5% ¹⁰Significant at 10%
¹¹Would have been significant at 11% level ¹²Would have been significant at 12% level ¹³Would have been significant at 13% level ¹⁴Would have been significant at 14% level ¹⁵Would have been significant at 15% level
Table 7: Firm specific exchange rate exposure to the Japanese Yen
Company name 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
HEINEKEN HLDG. -.2022¹⁰
UNILEVER CERTS. .2023¹⁰ -.1709¹⁴ .1340¹⁰
ROYAL DUTCH SHELL
A .1853⁵ -.4909¹ -.2283⁵ .1993⁵ .2891¹ BP -.6288¹ -.2073⁵ TOTAL -.7403¹ -.2154⁵ -.1259¹⁰ GAZPROM -137.3741⁵ -.3905¹ E ON .2542⁵ .3449¹ -.5832¹ -.3082⁵ ENI .2918⁵ -.1915¹⁰ 21.4339¹⁵ -.1304¹² GDF SUEZ -.3552¹ FIAT -.3918¹⁰ .4520⁵ SIEMENS -.2367¹ .2063¹⁰ -.2430¹⁰ .1572⁵ OIL COMPANY LUKOIL -.9921¹ -.2034¹⁵ -.4523¹ TESCO .3973¹ .1541¹⁰ ARCELORMITTAL -.5656¹ -.8199¹ METRO -.1573¹⁰ .2490¹⁰ EDF .3224⁵ .3108⁵ .1893¹³ TELEFONICA -.2088¹ -.1914⁵ PEUGEOT .4906¹ DEUTSCHE TELEKOM -.1371¹³ -.1533¹⁴ DEUTSCHE POST .2574¹⁰ .3808⁵ VODAFONE GROUP -.2440⁵ -.3530¹⁰ RWE .3102⁵ .3847⁵ -.2890⁵ .2245¹³ AIRBUS GROUP 18.4598 -123.7043⁵ .4769¹ .4049⁵ ORANGE -.2041¹⁵ -.2811¹¹ -.1896¹⁰ -.1997¹⁰ RIO TINTO -.2058¹⁴ -.8323¹ -.7489¹ -.2285¹⁰ A P MOLLER – -.5665¹ .1939¹⁴
MAERSK 'B' NOVARTIS 'R' .1802¹⁰ .1554¹⁰ SAINT GOBAIN .2527¹⁰ 20.6873¹⁰ NOKIA -1.0144¹⁰ -.2871¹¹ .2962¹⁰ VINCI -.2559¹³ .2333⁵ LYONDELLBASELL INDS.CL.A -.2339¹⁰ BAYER .3092¹² -4.3467¹⁴ SSE -.2402⁵ CASINO GUICHARD-P -.3428⁵ SANOFI -.3587¹ .1950¹⁰ VEOLIA ENVIRONMENT (OTC) .4994¹¹ VOLVO 'B' .3277¹⁰ -.3937¹⁰ -6.1621¹⁰ Overall result -8.5329¹
Table 8: Firm specific exchange rate exposure to the combined total currency returns factor
Company name 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
HEINEKEN HLDG. .0763¹⁰ .0808¹⁰ -16.4143⁵ .0789¹⁰
UNILEVER CERTS. .0806¹⁰ .1600¹ .1221¹ .0914¹⁰ .0856⁵
ROYAL DUTCH SHELL
A .1112¹ .0991¹⁰ -.1715¹ .1127¹ .1437¹ .2136¹ BP -.1179¹⁰ -.3206¹ -.1305¹ -.0657¹² TOTAL -.1467⁵ -.2078¹ -.0780¹⁰ .0733¹⁰ -.0514¹³ GAZPROM -.1513¹³ -.2188¹⁰ -.1258¹⁵ .1007¹⁴ -.1243¹⁰ E ON -45.4290¹¹ ENI -.1473⁵ -.3060¹ -.0868¹⁰ 12.2570¹⁰ -.0737¹¹ GDF SUEZ -.1728⁵ FIAT .1582¹⁵ .1799⁵ SIEMENS -.1376¹ .0747¹⁴ .1110¹ OIL COMPANY LUKOIL -.3193⁵ -.1526⁵ ENEL 15.5314¹³ TESCO .2177¹ .0820¹⁰ BASF -.1238¹⁵ ARCELORMITTAL -.3043¹ -.4634¹ NESTLE 'R' -.0832¹⁰ .0751¹⁰ .0637¹¹ METRO -.1976⁵ .1064¹⁴ .1555¹⁰ EDF . -.3700¹⁰ .0775¹² .0887¹³ .1511⁵ .1227¹⁰ TELEFONICA -.0842⁵ -.0633¹⁰ -.1256¹ PEUGEOT .1813 .1474¹² .1283¹⁰ -.2392¹³ DEUTSCHE TELEKOM -.0949⁵ 7.6925¹¹ DEUTSCHE POST .1349¹⁵ -.1806¹ -12.6265¹¹ .0844¹⁰ .1067¹⁰ VODAFONE GROUP -.1483¹ THYSSENKRUPP -.1026¹⁰ -.2036¹ -.1493¹¹ -.1290¹¹ RWE -.1193⁵ -.1922¹
AIRBUS GROUP -40.6726¹² .2287¹ .2668¹ .2456¹ .1552¹⁵ ORANGE -.1428¹⁰ -.0791¹¹ -.0886¹⁰ RIO TINTO -.1373⁵ -.2598¹ -.4690¹ -.4213⁵ -4369¹ -.1838¹⁰ -.1638¹¹ -.2039¹ A P MOLLER – MAERSK 'B' -.2708¹ NOVARTIS 'R' .1474¹ .0342⁵ .1197¹ SAINT GOBAIN .1010¹³ 11.8503⁵ .1287¹¹ NOKIA -3.7394¹⁴ -.1478¹⁵ -27.9940¹⁰ .2061⁵ VINCI .1232⁵ -.0656¹¹ BAYER .0962¹⁴ ROCHE HOLDING .0886¹⁰ CASINO GUICHARD-P -.1231¹⁰ -.0981¹³ .0914¹⁰ .0907¹² SANOFI .1036¹¹ .0100¹² -.1440⁵ .0739¹⁰ .0665¹⁵ .1040¹⁰ VOLVO 'B' .1185¹³ -.1753¹⁰ -.3414¹ Overall result -3.8414¹ .2999¹⁵ .7017¹⁰ Table 9: Summary statistics of FCD and currency factors Table 10: Correlations FCD to currency factors
Table 11: Regression FCD on currency factors
