Master Thesis Finance Foreign Currency Exposure, Financial Hedging Instruments and Firm Value

Master Thesis Finance 2012 Foreign Currency Exposure, Financial Hedging Instruments and Firm Value Author : P.N.G Tobing Student number : U1246193 ANR : 187708 Department : Finance Supervisor : Dr.M.F.Penas Faculty name : School of Economics and Management Year of graduation : 2012 Word count : 13308 1 P a g e

Abstract This paper examines whether the foreign currency exposure influences firm s choice among financial hedging instruments. This paper used various measures of currency exposure with different characteristics in order to distinguish between each financial hedging instrument. By using newer dataset that consists of 188 manufacturing firms in the US during 2006, this paper finds that various proxy of currency exposure is positively and significantly associated with a particular type of financial hedging instruments. Further, the results also show that each financial hedging instrument is a complement to one another in reducing the currency risk exposure. In addition, by differentiating type of financial hedging instrument this paper is also able to investigate the effect of hedging on firm value in clearer ways. After the inclusion of foreign debt in the definition of financial hedging activity, this paper does not find evidence that financial hedging increase firm value. 2 P a g e

Table of Contents Abstract... 2 1. Introduction... 4 1.1 Background... 4 1.2 Research questions... 7 1.3 Structure of the thesis... 8 2. Foreign currency exposure... 9 3. Hedging behavior... 11 3.1 Rationale of hedging behavior... 11 3.2 Financial hedging instruments... 13 3.3 Valuation effect of hedging... 18 4. Sample selection and methodology... 20 4.1 Sample selection... 20 4.2 Methodology... 23 4.2.1 Foreign currency exposure and financial hedging instruments... 23 4.2.2 Relationship between financial hedging instruments... 25 4.2.3 Financial hedging and firm value... 25 4. 3 Variable construction... 26 4.3.1 Financial hedging instruments... 26 4.3.2 Foreign currency exposure... 30 4.3.3 Firm value... 30 4.3.4 Control variables... 31 5. Empirical results... 35 5.1 Logit regression result... 35 5.2 Multinominal logit regression result... 41 5.3 OLS regression... 45 6. Conclusion... 51 7. References... 54 3 P a g e

1. Introduction 1.1 Background A substantial amount of theories for optimal hedging have been developed based on the Modigliani and Mihedgingller paradigm. Those optimal hedging theories provide explanation for hedging behavior based on the capital market imperfection. Various earlier empirical researches have examined the determinant of foreign currency hedging, but the focus of the earlier empirical researches are on whether the usage of derivatives conforms with the managerial risk aversion, financial distress, tax liability and underinvestment problem theories discussed by Mayers & Smith (1990), Smith & Stulz (1985) and Froot, Stein & Scharfstein (1993). Moreover, earlier researches also limited the sample by taking data during 1990s (Allayannis & Ofek, 2001; Geczy, Minton & Schrand, 1997; Graham & Rogers, 2002; Mian, 1996), where only limited amount of studies have analyzed the hedging strategies by using more recent data. This paper addresses this gap, by using newer dataset to analyze the financial hedging activity in US. The sample consists of 188 manufacturing firms in the US during 2006. FASB issued Statement no. 133, Accounting for Derivative Instruments and Hedging Activities in June 1998 that is mandatory for fiscal years beginning 2000. Due to the implementation of FASB 133 disclosure for hedging and derivative instrument in 2006 become more transparent and as a result facilitates the hand collected data method used by this paper. This paper argues there are other factors besides those that conform to the optimal hedging theory, which also influence the firm hedging behavior. This argument is based on another stream of theory developed by Ederington (1979). He argue that a firm s decision to hedge the exposure reduce as the firm face greater basis risk. Basis risk is defined as the difference between the characteristics of exposure being hedged and characteristics of hedging instrument (Ederington, 1979) Therefore, in the context of foreign currency hedging activity, this paper argues that the foreign currency exposure itself that influences the choice among hedging instruments. Size, frequency and timing of the exposure are example of important currency exposure characteristics that must be consider in order to choose a particular hedging instrument. Therefore, consistent with Ederington (1979), this paper argue that the higher the correlation 4 P a g e

between the source of exposure and the hedging instrument, the lower the basis risk faced by the firm and as a result the probability of using particular hedging instrument also increase. Similar research has been made by Clark & Judge (2009) in the sample of UK non-financial firms during 1995. Clark & Judge (2009) provided evidence that the currency exposure is an important determinant for currency hedging activity in the UK. The motivation for conducting similar research in US is that there are several major differences in the firm characteristic among the US and UK firms that might influence the firm hedging activity in each country. The short liquidity factor (measured by quick ratio and dividend yield ratio) and growth opportunities factor (measure by RnD expense/total assets) highlighted those major differences in characteristics between the US and UK firms. The short term liquidity factor captures the availability of firm internal fund, while RnD ratio indicates firm growth opportunities. The summary statistic of this paper shows that the quick ratio for firm in the US sample is above the sample mean reported by Clark & Judge (2009). Consistent with this, the dividend yield ratio for firm in the US sample is also below the dividend yield ratio in the UK sample reported by Clark & Judge (2009). The growth opportunities on average is similar to those reported by Allayannis & Weston (2001) and Elliot, Huffman & Makar (2003) in the context of US firm, but the ratio is far below the number reported by Clark & Judge (2009) in the sample of UK firms. Therefore although similar research have been made in UK, it is still interesting to see whether the currency exposure is an important determinant of hedging behavior in the context of US firms. The data related to the use of derivatives is published by the Bank of International Settlements and BIS (2007) shows that the notional amount outstanding of over-the-counter derivatives increased from US$257.9 billion in December 2004 to US$415.8 billion in the same month in 2006 (BIS, 2007). Further, there percentage of hedger in US from 1990 to 2006 also increases. Geczy et al. (1997) show that more than half of firms in 1990 are non-hedgers, while in 2006 this paper report that only 18 % of firm in the US sample is non-hedgers. Interestingly, Clark & Judge (2009) report that 30 % of firm in the UK sample is non-hedgers. The growing number of hedgers in the US during 2006 is somewhat contradict the prediction of earlier optimal hedging theory. Consistent with the cost of financial distress and the underinvestment problem theories, 5 P a g e

higher liquidity factor and lower growth opportunities reduce the need for hedging, while in the reality the currency hedging activity in the US is going in the opposite direction. This contradiction also provides support for the earlier argument of this thesis that there are other factors, besides those that conform to the theory prediction, that influence the firm hedging behavior. Therefore it is interesting to see in multivariate setting analysis, whether the currency exposure itself that influences a firm choice s among hedging instruments. By analyzing the relationship between the foreign currency exposure and choices among different type of hedging instruments, this paper is also able to determine the relationship between each type of financial hedging instrument; whether they act as complement or substitute to one another. If a type of exposure is an important determinant for a particular type of instrument, therefore each instrument is unique to a particular exposure, providing evidence for the complementary relationship between each financial hedging instrument. On the contrary if a particular type of exposure is an important determinant for various hedging instrument, then each financial hedging instrument is a substitute to one another. Therefore, the result of this paper extend the work of Allayannis & Ofek (2001) and Geczy et al. (1997) that examined the determinant of foreign currency hedging activity based only on the optimal hedging theory. Further, by differentiating type of financial hedging instrument this paper is also able to investigate the effect of hedging on firm value in clearer ways. Various studies have investigated the valuation effect of hedging by examining the relationship between foreign currency derivative (FCD) and firm value. Majority of the existing studies find positive relationship between FCD and firm value (Allayannis & Weston, 2001; Carther, Rogers & Simkins, 2006). The more recent studies made by Allayannis, Lel & Miller (2011) and find that the usage of FCD increase firm value only for firm that have strong corporate governance. Overall, in examining the valuation effect of hedging, the focus of the earlier empirical researches is on the FCD. Only limited amount of studies are known for investigating the valuation effect of foreign debt (FD). This paper argues that the positive relationship between hedging and firm value is due to what is included in the definition of financial hedging. The inclusion of FD user in the hedging definition may change the valuation result of financial hedging activity and further, challenge the result made Allayannis et al. (2011), Allayannis & Weston (2001) and Carther et al. (2006). Therefore, 6 P a g e

this paper raises another interesting research question on whether the usage of FCD in combination with FD will contribute differently to firm value. 1.2 Research questions Instead of looking at the foreign currency derivatives (FCD) alone this thesis also extends the definition of financial hedging by including the foreign debt (FD) users. The main question to be addressed in this thesis is whether the currency exposure influences firm s choice among hedging instruments. By analyzing the relationship between the currency exposure and choice among different types of hedging instruments, this paper hopes to answer the question as to whether each financial hedging instrument act as a complement or substitute to one another. Furthermore, by differentiating the types of financial hedging instruments, this paper addresses another interesting research question whether the financial hedging activity contributes to firm value, even after the inclusion of FD user in the financial hedging definition. To test the first and second hypotheses, this paper uses logit regression model that is similar to Allayannis & Ofek (2001). Further in order to investigate the relationship between each hedging instrument, this paper also uses the multinomial logit regression model similar to Clark & Judge (2009) and Geczy et al. (1997). The logit and multinominal logit regression show the probability of a firm using a particular hedging instrument given its exposure factor. Therefore, each of the hedging instruments will be regressed to the various proxy of foreign currency exposure also by controlling for other variables that influence hedging activity. For the firm value regression, this paper will use OLS basic model similarly used in Allayannis, Ihrig & Weston (2001), Allayannis et al. (2011), Allayannis & Weston (2001), Carther et al. (2006) and Kim, Mathur & Nam (2006) who regressed the firm value with hedging activities in the multivariate settings analysis. Similar to those earlier studies, Tobin s q will be used as proxy for firm s market value. Overall, by using sample of U.S multinational manufacturing firms during 2006, this paper finds several result. Firstly, this paper shows that currency exposures is an important determinant for a firm s choice among financial hedging instruments, consistent with the basis risk theory and provide supporting evidence for the hypotheses. Secondly, the paper also shows that each financial hedging method is a complement to one another in reducing the currency risk as the 7 P a g e

usage of a particular financial hedging method is appropriate for a particular type of currency exposure. Thirdly, the paper shows that after the inclusion of FD, the financial hedging activity do not increase firm value. Interestingly, even after this paper control for the usage of foreign debt, the result provides no evidence that the usage of FCD alone increase value of the firm. These findings contradict the earlier empirical studies made by Allayannis et al. (2001), Allayannis et al. (2011), Allayannis & Weston (2001), Carther et al. (2006) and Kim et al. (2006). One possible explanation for this is that those manufacturing firms are multinationals that are diversified geographically. Firms that are diversified geographically are more likely to match the revenue and cost, as a result the level of currency exposure reduces as the firms are more geographically diversified. Therefore the usage of financial hedging is not found to be beneficial to firm value. 1.3 Structure of the thesis The organization of this thesis is as follows: Chapter 2 discusses the different characteristic of foreign currency exposure. Chapter 3 discusses the hedging behavior based on existing theories and literature. Chapter 4 provides the sample selection criteria, methodology used for testing the hypotheses, as well as the summary statistic and pearson correlation coefficient. Chapter 5 presents the result of the empirical research of this paper and the answers for the research question. Chapter 6 provide provides concluding remarks on the thesis as well as the contributions of this paper for organizations as a whole, annual report users as well as the accounting standard bodies that set disclosure requirement for hedging, derivative and financial instruments. Lastly various limitations of this paper and opportunities for future research will also be provided in the last chapter. 8 P a g e

2. Foreign currency exposure Ederington (1979) and Haushalter (2000) emphasize the importance of determining the characteristics of each currency exposure. The motivation for this argument is that each currency exposure has specific characteristic that eventually impact the usefulness of each financial hedging instrument in reducing that particular exposure. For instance the exposure arising from foreign debt is generally long term in nature and requires multiple subsequent payments from the determination until the maturity of the contract. Further, the size and the timing of the payment also depends on the initial agreement between the two parties (pre-determined), therefore the exposure is characterized by multiple and certain transactions. On the other hand, currency exposure resulted from foreign transaction such as purchase of raw materials are considered to be uncertain with respect to size and timing, as the purchase quantity, price per unit and timing of installment varies for each purchase transaction. A firm s purchase of raw material is also considered to be a short term transaction as it generally requires a single subsequent payment or in the context of payment on credit (e.g. trade payable account balance) the installment generally settles within one year. Therefore although the purchase transaction is frequent but the transaction is also uncertain with respect to amount and timing. By determining the characteristics of each exposure, firms are able to match the characteristics of the exposure being hedged with the financial hedging instrument used to hedge. The motivation is based on the basis risk argument developed by Ederington (1979). He argues that a firm s decision to hedge exposure reduces as the firm faces greater basis risk from using a particular instrument. Basis risk is defined as the difference between characteristic of the underlying currency exposure being hedged and characteristic of underlying hedging instrument such as maturity and price (Ederington, 1979). The lower the correlation between hedging instrument and source of exposure, the higher the basis risk faced by the firms, therefore the less useful is hedging instrument, consequently the less extensive firm should hedge the exposure. Therefore this paper argues that firm should identify all components of their business that affected by the exchange rate movement and understand the characteristics of each exposure as the exposure factor is an important determinant of firm s choice for particular hedging instruments. Later on 9 P a g e

this paper use various measure of currency exposure with different characteristic in order to distinguish between each financial hedging instrument. Hypothesis 1: The foreign currency exposure affects firm choice among hedging instrument. 10 P a g e

3. Hedging behavior This chapter discusses the firm s hedging behavior based on existing theories and literature. Section 1 discusses the rationale for hedging behavior based on existing theories. Section 2 discusses the type of financial hedging instruments and reviews progress of previous studies that investigated the relationship between each financial hedging instruments. Section 3 discusses previous studies that examined the linkage of hedging and firm value as well as the potential channels through which hedging may affect differently on firm value. 3.1 Rationale of hedging behavior For the purpose of this research hedging refers to the usage of financial instruments or restructuring business activities to create natural hedge, both activities aim at reducing or eliminating adverse effect on cash flow due to exchange rate movement (Adler & Dumas, 1984). Optimal hedging theories rely on the capital market imperfection to explain firm incentive to hedge. In these theories, the hedging activity employed by the firm is assumed as a strategy to reduce the variability of cash flow. Smith & Stulz (1985) argue the tax liability theory. The argument is based on the firm ability to reduce the variability in income via hedging and as a result the expected tax liability born by the firm is also reducing. Empirical studies find mixed results regarding to the relationship between hedging and tax reduction (Graham & Rogers, 2002; Nance, Smith & Smithson, 1993). Managerial risk aversion theory looks at the impact of managerial wealth and managerial stock option on corporate hedging decision (Mayers & Smith, 1990; Smith & Stulz, 1985) The probability of hedging increases as the manager owns large number of firm shares and in contrast the probability of hedging decreases as the manager own more stock options. Empirical research such as Schrand, Catherine & Unal (1998) and Tufano (1996) find supporting evidence for the managerial risk aversion theory, but majority of the empirical research such as Allayannis & Ofek (2001), Geczy et al. (1997), Graham & Rogers (2002) and Haushalter (2000) find no evidence showing that managerial risk aversion affect the corporate hedging decision. Smith & Stulz (1985) also point out the cost of financial distress theory. A firm with higher probability of financial distress is more likely to hedge as hedging decreases the variance in firm 11 P a g e

cash flow. Empirical studies on this theory use debt ratio as a proxy of financial distress and quick ratio and dividend yield as proxies for short term liquidity. Bartram, Brown & Fehle (2004), Dolde (1995) and Graham & Rogers (2002) find evidence that showing hedging increases as firms have more leverage and lower liquidity, thus support the cost of financial distress theory. While, Nance et al. (1993) find no supporting evidence showing that hedging increase as leverage increase. Further they argue that the insignificant relationship between leverage and hedging indicate that other than hedging, a firm is also able to reduce the probability of financial distress by maintaining higher liquidity and lower dividend yield. Moreover, Smith & Stulz (1985) also argue that by decreasing the probability of distress, hedging raise the ability of firm to obtain external finance. As a result the probability of firm to underinvest is also decrease. In conjunction with this, Froot et al. (1993) argue that by reducing variance in cash flow, hedging also reduces firm dependency on external finance. Thus, in the situation where external finance is costly, firm is able to take the investment opportunities as hedging help firm to maintain sufficient internal fund. Explanatory variables such as size and investment opportunities are used to explain the relationship between firm characteristics and hedging (Dolde, 1995; Geczy et al., 1997; Mian, 1996). Majority of the empirical studies find a positive relationship between hedging and proxies of investment opportunities such as RnD and Capex spending (Geczy et al. 1997), thus supporting the notion that hedging helps companies to maintain adequate funds available for good investment opportunities. Other stream of research also argues that size is as an important determinant of corporate hedging. The cost of entering and maintaining FCD might be too high for small firms, therefore limiting their ability to use FCD (Allayannis & Ofek, 2001; Geczy et al., 1997; Mian, 1996; Tufano, 1996). Therefore, hedging activity decreases for smaller firms and increases for bigger firms. However, consistent with the underinvestment theory, Froot et al. (1993) also provide supporting argument that smaller firms are more likely to employ hedging activity as smaller firms have higher expected growth compare to the larger firms which generally are more mature thus have lower growth opportunities. Overall, the consensus among those earlier theories and empirical researches is that they largely focus on understanding why firm s hedge. This paper argues that those existing empirical 12 P a g e

researches overlook the optimal hedging theory. Furthermore, there are other factors such as the currency exposure, which is also an important determinant of firm hedging decision. The basis risk theory discussed earlier in Chapter 2 motivates this argument. 3.2 Financial hedging instruments Despite focusing on understanding why firm s hedge, the existing empirical studies of currency hedging has also focused on the usage of foreign currency derivatives (FCD) with no distinction on the different types of hedging method employed by firm. In this section this paper discusses different type of financial hedging methods that will be used in this paper; short term financial hedging in one hand that include the usage of FCD such as forwards and options as well as swap and foreign debt on the other 1. Among all instruments, forward and options are the most popular instrument used in practice to hedge the currency exposure. Geczy et al. (1997) examine the FCD usage for 372 firms in US during 1990 and finds that 29.3 % of firms in the sample are forwards users. Clark & Judge (2009) finds that among 412 firms in the UK during 1995, 48.4 % of the firms use forwards only and 22 % of firms use forward in combination with other instrument. Risk management literatures state that forwards protect the holder from unfavorable changes in exchange rate but do not permit the holder to obtain the benefit from the favorable movement in exchange rate, while options is more expensive than the other derivative instruments, as the options feature permit the holder to benefit from favorable movement in exchange rate (Sundaram & Das, 2011). Various studies on currency hedging in the past have examined the usage of FCD but the majority of them tend to exclude the usage of currency swap (Allayannis & Weston, 2001; Bodnar & Gentry, 1993). The reason is due to the characteristics of swap that is not consistent with forward. In swap the two parties exchange fund directly rather than traded with banks or other OTC market. Further, Geczy et al. (1997) argue that in swap both parties are able to customize the contract based on their needs. Forward allows the firms to take opposite position to its spot market position by buying or selling a specified amount of a currency at a predetermined rate on a particular date in 1 None of the firm in the sample provide any indication of future usage, therefore future contract is excluded from the scope of this study 13 P a g e

the future. Therefore, similar to swap, forwards can be customized to meet the needs of the firms, but the maturity is seldom above one year. Therefore, the customization cost is relatively lower for single long term contract (swap) than for series of short term contract (forward). Therefore, risk management literatures argue that swaps provide long term flexible hedge with relatively low transaction cost. Other earlier studies also argue that the exclusion of swap in the regression do not provide significant difference to the result as there are very limited number of firms used swap to hedge currency risk (Allayannis & Weston, 2001; Bodnar et al., 1993). However, Clark and Judge (2009) and Elliot et al. (2003) find that currency swaps is used by those firms that have foreign denominated debt. JOHNSON & JOHNSON in its 2006 annual report stated that: The Company uses forward exchange contracts to manage its exposure to the variability of cash flows, primarily related to the foreign exchange rate changes of future intercompany product and third party purchases of raw materials denominated in foreign currency. The Company also uses currency swaps to manage currency risk primarily related to borrowings. Both of these types of derivatives are designated as cash flow hedges (p. 56). Therefore, the quote above supports the importance of swaps and the inclusion of that instrument in the definition of financial hedging method employed by firm. Among all studies in currency hedging, Geczy et al. (1997) specifically examine the usage of FCD and the choice of hedging instruments in the sample of US firms during 1990. They argue that the source of exposure affect the level of benefit that can be realized from hedging. However, Geczy et al. (1997) limit the definition of hedging into the usage of derivatives and as a result they differentiate between the forwards and options only users with the swap only users and exclude the users of foreign currency debt (FD). In a univariate analysis, Geczy et al. (1997) find that currency swaps are more cost-effective for hedging foreign debt risk, while forward contracts are more cost-effective for hedging foreign transaction risk. The reason is due to the notion that foreign debt payment fits the long-term nature of currency swap contract. While the frequency and uncertainty of transactional exposure need to be managed dynamically by short term forward contract. However, in the multivariate setting they still find mixed results on the 14 P a g e

relationship between currency exposure and the choice among financial currency derivative instruments. The study on natural hedging such as foreign debt (FD) has received less attention. In examining the choice of financial hedging instruments, this paper argues that the exclusion of foreign currency debt in the definition of financial hedging activity might bias the result, as a firm might be classify as non-hedgers when it uses FD to hedge. This argument based on the evidence found by Allayannis & Ofek (2001) and Elliot et al. (2003) showing that FD is also used as a hedging strategy in mitigating currency risk exposure. The usage of FD illustrated by the following example, an exporter who will received revenue in foreign currency bearing the risk that the foreign currency will be depreciate against the domestic currency. Therefore, the exporter is able take a loan in foreign currency and converts the loan to domestic currency. By assuming that the gain realized by investing the proceeds from the loan will match the interest rate payment, at the maturity date the foreign denominated loan will be payback with the amount from foreign revenue. Therefore, FD creates stream of cash outflow in a foreign currency that match the company foreign revenue. Hagelin & Pramborg (2004) for a sample of Swedish firm provide supporting evidence that FD is an effective tool for risk reduction. Therefore, if the derivatives market is too expensive, a company has the alternative of using a FD, providing evidence that FD is a substitute of FCD in reducing the currency risk (substitution hypothesis). If each hedging instrument have its own characteristics as describe above, then each instrument can be match with the characteristics of each exposure, thus currency exposure is an important determinant for corporate hedging activity as the firm s choice among a particular hedging instruments is depend of the types of exposure (Hypothesis 1). Furthermore, the matching concept also raises another argument that each financial instrument is also appropriate hedge certain type of exposure consistent with the basis risk theory describe earlier (Chapter 2) and provide supporting argument for the complementary relationship between each financial hedging instruments. For instance, exposure arising from possession of asset in foreign countries is obviously a long term exposure. Mismatch between the duration of the foreign asset and hedging instrument, will increase the basis risk discussed earlier. Therefore, the basis risk from using short term financial hedging instrument will be higher relative to the usage of swap or FD. Swap 15 P a g e

and FD are relatively a low cost method for hedging the long term type of exposure, provide evidence for the complementary relationship between the usage of forward and options and the usage of Swap or FD. Another stream of studies has attempted to directly examine the relationship between FD and FCD by incorporating FD into the analysis of FCD. Fok, Carroll & Chiou (1997) use the logit model to find evidence showing that convertible debt serves as substitute for derivative used. Allayannis & Ofek (2001) use the logit model and test a model of choice between FCD and FD of U.S non-financial firm. They find evidence showing that exporters prefer to use FCD than FD but do not find significant evidence multinationals prefer to use FD over FCD. However, their results might be limited due to the exclusion of swap. Bratham et al. (2004) include swap in the definition of financial derivatives and find a positive relationship between FCD and FD, provide evidence on the complementary hypothesis. Elliot et al. (2003) examine the relationship of FD and FCD for sample of U.S MNC during 1994-1997. Unlike other studies that use binary variable as measure for FCD and FD usage, this study uses continuous variable. By using the national value of FCD as a proxy for FCD and book value of FD as a proxy for FD usage, Elliot et al. (2003) find a negative relationship between FD and FCD. This result implies that both FD and FCD used to hedge the currency risk and FD used interchangeably with FCD to hedge the currency risk. The result is consistent with Allayannis & Ofek (2001), Fok et al. (1997) and Hagelin & Pramborg (2004) and provides evidence to support the substitution hypothesis. Clark & Judge (2009) in the sample of UK firms during 1995 investigate the relationship between FCD and FD by examining whether a particular financial instrument is unique to a particular exposure or associated with all type of exposure. In the research, Clark & Judge (2009) distinguish swap from other financial derivatives such as forward and options and further divide the financial hedging become two elements; the usage of forwards, futures and options that classified as short term FCD on one hand and the usage of FD or swap (independently or in combination with short term FCD) that classified as long term financial hedging method on the other hand. This classification is based on the characteristics of swap that are more similar to foreign debt which are long term in nature. For example PROCTER & GAMBLE CO in its 2006 annual report supported the use of swap and FD to hedge the long term currency risk and stated 16 P a g e

that: We hedge certain net investment positions in major foreign subsidiaries. To accomplish this, we either borrow directly in foreign currency or designate all or a portion of foreign currency debt as a hedge of the applicable net investment position or enter into foreign currency swaps that are designated as hedges of our related foreign net investments (p. 53). By using multinominal logit regression model, Clark & Judge (2009) find evidence for the complementary hypothesis. In examining the relationship between each financial hedging instrument, this paper will follow Clark & Judge (2009) that regress each of the hedging instrument with various measure of currency exposure and investigate whether each hedging method is specific to hedge a particular or appropriate to hedge all types of exposure. However, instead of distinguishing financial hedging become two elements, this paper distinguish firm that use forwards, futures and options only, firm that disclose the use FD only and firm that disclose the use swap (independently or in combination). The intuition behind this argument is that consistent with the basis risk theory, a particular instrument is unique to a type of exposure, therefore each hedging instruments is a complement to one another. But if all instruments can be used interchangeably to hedge all types of exposure, then each hedging instrument is a substitute to one another, therefore support the earlier empirical results made by Allayannis & Ofek (2001), Elliot et al. (2003), Fok et al. (1997) and Hagelin & Pramborg (2004). Hypothesis 2: If each source of foreign currency exposure is positively related to a particular type of hedging instrument then each financial hedging instrument is a complement to one another in reducing the currency risk exposure. But if a particular type of exposure is an important determinant for various hedging instrument, then each financial hedging instrument is a substitute to one another in reducing currency risk exposure. 17 P a g e

3.3 Valuation effect of hedging Given the fact that companies make extensive use of various hedging methods raises further question on the valuation effect of those hedging methods employed by the firms; whether or not financial hedging method employed by the firm contribute to the firm value. Various studies have examined directly the valuation effect of hedging strategies on firm value. Allayannis & Weston (2001) who use a linear model for the U.S. data, find the usage of FCD increases total firm value as much as 4.8 percent on average. Further, Nain (2004) also show that FCD increases value of the firm by 5% on average. Moreover, he also shows that a firm that remained unhedge will suffer a value discount if the competitors in the industry hedge their foreign currency exposure. Allayannis et al. (2001) find that operational hedging increases firm value when combined with FCD, while Kim et al. (2006) find that both operational and FCD strategies are positively associated with firm value. Operational hedging adds 4.8-17.9% on average, while financial hedging adds 5% on average to firm value. Interestingly, Bartram et al. (2004) examine large sample of multi industry companies from 48 countries. They find that although the majority of derivative user is dominated by FCD users (35.9%), the use of interest rate derivatives (not the currency derivatives) that associated with firm value. Therefore, there are no value effect of FCD. Furthermore, Allayannis et al. (2011) find that the hedging premium is statistically significant and economically large only for firms that have strong internal and external corporate governance. Clark & Judge (2009) using sample of UK non-financial firms show that different hedging strategies make different impact to the firm value. When hedging strategy broken down into FCD and FD user, the usage of FCD is associated with higher firm value by 14% on average. However, FD is associated with higher firm value only when the usage of FD is combined with FCD. Interestingly while other studies exclude swap from the definition of FCD, Clark & Judge (2009) find that the usage of swap is found to be more effective in increasing firm value than FD and any FCD instruments. 18 P a g e

Overall, there are still mixed results on the valuation effect of hedging. Furthermore, the focus of the earlier empirical researches are on the valuation effect of FCD and there are only limited amount of studies investigate the valuation effect of FD. Similar to those UK paper, this paper provides clearer test on the valuation effect of hedging in the context of US firm by looking at the impact of each type of hedging instrument specifically the role of FCD in combination of FD in increasing firm value. The intuition behind differentiating types of financial hedging instrument is that the effect of hedging and firm value is influence by what is included in the definition of financial hedging. Therefore, this paper argues that the inclusion of FD in the definition of financial hedging activity may provide different result from the earlier empirical paper and contradict the result made by Allayannis et al. (2001), Allayannis et al. (2011), Allayannis & Weston (2001) and Kim et al. (2006). Hypothesis 3: A positive and significant relationship between each hedging instrument and firm value would indicate that hedging increases firm value and a negative and significant relationship between each hedging instrument and firm value would indicate that hedging decreases firm value.. 19 P a g e

4. Sample selection and methodology The first section explains the sample selection criteria and summary statistic for firms in the sample. The second section explains the econometrics methodology. The third section presents description of the variables used in the analysis as well as the correlation between variables. 4.1 Sample selection Following Kim et al. (2006), the initial sample is obtained from Compustat Geographic Segment files for year 2006. The firm included in the sample must be non-oil related manufacturing industry, thus this paper only include those firms that within 2000-2899 or 3000-3999 SIC codes range (Kim et al., 2006). There are 482 firms with foreign sales or export sales within those SIC codes range. Export sales are defined as the revenue generated from production domestically but sold outside the U.S (Kim et al., 2006). This paper only focuses on firm that the main (domestic) operation is in the U.S area, which reduces the sample size into 340 firms. Further, firm in the sample must also be exposed to foreign involvement above 20% (Elliot et al., 2003) and the size of the firm is above $500 million (Allayannis & Weston, 2001). Finally the sample is also restricted to December year end in order to facilitate consistent assumption about the exchange rate (Guay and Kothari, 2003). Therefore the sample reduces into 200 firms. After the size, industry and foreign exposure matching requirements, only a firm that has 10- K/annual report available from EDGAR Database is included in the sample. Further the data on market value must be also available from the Datastream. After those selection criteria, the final sample consists of 188 US manufacturing firms. Table 1 provides the summary statistics for the sample of 188 US multinationals firms during 2006. Panel A shows the international characteristics of the firms. The foreign sales ratio (Fsales) proxy for currency exposure arising from possession of foreign assets ranges from 30% to 98%, with mean (median) of 52% (49%).This percentage is greater than the percentage reported by the earlier researches such as Allayannis & Weston (2001) and Geczy et al. (1997) in the sample of US firms and further, Clark & Judge (2009) in the sample of UK firms, that indicate the mean level of foreign sales is 18% and 38% of total sales for the sample of US non-financial firms and 35% of total sales for the sample of UK non-financial firms. However, this percentage is slightly 20 P a g e

above the number reporter by Elliot et al. (2003) that indicate the exposure through foreign sales for U.S multinationals from 1995 to 1997 is 44.3% on average. Therefore, the firm in the sample exposed to higher foreign currency exposure as compare to other earlier US and UK firms. One possible explanation is that firms in 2006 are more engage in the international activities compare to firms in the 1990s. Panel B shows the general firm characteristics. The mean (media) book value of total assets is $10.18 billion ($2.5 biilion). The total assets are below the number reporter by Elliot et al. (2003) that indicate the total asset is $ 14 billion ($3.7 biilion). But this number is greater than reported by Allayannis & Weston (2001) with $7.7 million ($2.5 million). The mean market value of equity is $13.47 billion. Tobin s Q is the proxy of firm value. The first measure of Tobin s Q is book value of total assets minus the book value of equity plus market value of equity divided by the book value of total assets for the year ended 2006. The median value (0.54) of Tobin s Q is smaller than the mean (0.60), which indicate that the distribution of Tobin s Q in the sample is skewed. Thus, this paper use natural log of Tobin s Q to control for this. The mean (median) for quick ratio is 76% (40%) which is above the sample mean (media) by Clark & Judge (2009) with 48% (30%), although that the mean of long term debt to total assets ratio (Tdebt) is 20%, which is slightly higher than reported by Clark & Judge (2009) and Elliot et al. (2003) with 18% and 17% on average. This is an interesting point as although the US firm is less financially flexible due to the higher debt ratio, but on average firm in the sample is cash rich company, which unnecessary to have foreign borrowing. The higher cash holding status is also consistent with the relatively low dividend yield ratio of 1 % on average which is extremely below the dividend yield ratio in the UK sample reported by Clark & Judge (2009) of 3.5% on average. Thus, given its high cash status, the majority of firm might have preference on FCD instead of foreign debt to hedge the risk; consistent with the argument that firm with higher level of liquidity might have greater ability for using foreign currency derivatives (Allayannis & Ofek, 2001; Geczy et al., 1997; Mian, 1996; Nance et al,. 1993; Tufano, 1996). However, the 21 P a g e

relationship between firm liquidity and derivative usage can also be negative. According to the financial distress theory (Smith & Stulz, 1985) firm with greater probability of financial distress is more likely to use derivatives, thus the higher quick ratio and lower dividend yield ratio reduce the need to use derivative to reduce the expected cost of financial distress (Smith & Stulz, 1985) and external financing (Froot el at., 1993). Further, the growth opportunities measured by RnD expenses divided by total assets is 4% on average similar to reported by Elliot et al. (2003) with 4.4 % on average and Allayannis & Weston (2001) with 3% on average, although that this ratio is far below the number reported by Clark & Judge (2009) with 80.4%. This is also another difference between firm characteristic in US and UK. According to Froot el at. (1993) hedging should be greater for firm with higher investment opportunities to reduce the underinvestment problems. Based on above data, UK firms have extremely higher investment opportunities compare to US. Thus, consistent with the underinvestment theory Froot el at. (1993), the usage of derivative in the US is expected to be lower compare from the UK. Table 1: Summary statistics This table provides summary information for all variables used in the regression. Variable Mean Median sd max min N Panel A: Firm international characteristics STexp.85 1.36 1 0 188 Fsales (%).52.49.15.98.30 188 Netinvestm~k.82 1.38 1 0 188 FDrisk.21 0.41 1 0 188 Panel B: General firm characteristics AT (million) 10180.31 2520.34 25405.45 278554 511.603 188 MVE (million) 13478.69 2767.68 29618.01 191011 41.8552 188 Tobin s Q (LNQ1).60.54.39 2.06 -.32 188 Leverage.20.13.19 1.26 0 188 Tdebt.20.16.18 1.61 0 188 DivDummy.53 1.49 1 0 188 DivYield.01 0.01.04 0 188 Profitabil~y.14.13.07.49 -.03 188 GrowthOpp.04.03.03.20 0 188 TaxDummy.86 1.34 1 0 188 Quickratio.76.40 1.01 7.18.012 188 22 P a g e

4.2 Methodology This section aims to construct empirical model in order to test each of the hypotheses mentioned in Chapter 2. Firstly, this section explains the methodology use to test whether the currency exposure is an important determinant of hedging activity and secondly whether each particular exposure specific for a particular instrument or appropriate for all types of financial hedging instruments in order to determine the relationship between each financial instrument. Lastly, this section explains methodology use to test the valuation effect of hedging. 4.2.1 Foreign currency exposure and financial hedging instruments This paper divides the financial hedging instrument users become three categories: the user of short term foreign currency derivatives (SHFCD), the user of cross currency swap (SWAP), and the user of foreign debt (FD). The grouping is designed to differentiate types of hedging instrument based on its characteristics such as maturity, degree of customization and cost. The grouping is similar with Clark & Judge (2009) that distinguish forwards, futures and options users and long term foreign currency hedging users that include swap or FD users. Therefore, the dependent variable is those three types of financial hedging instrument categories (Finhedge). The first dependent variable is the SHFCD dummy that equal to 1 if firm indicate any forwards or options usage and 0 otherwise. The second dependent variable is SWAP dummy that equal to 1 if firm indicate any swap usage and 0 otherwise. The third dependent variable is FD dummy that equal to 1 if firm indicate any foreign debt usage for hedging purpose and 0 otherwise. For each of these hedging instrument categories, this paper examine whether the different measure of currency exposure are important determinant of financial hedging instrument. The most common method for analyzing data with binary response variable is the logit regression method. The logit model in this paper is use to examine whether the various measure of currency exposure are important determinant of currency hedging activity. Further, this logit model is also use to estimate the probability of using a particular hedging instrument given its exposure type. This logit model is similar to the model used by Allayannis & Ofek (2001) that examine the determinant of FCD based on firm characteristics. 23 P a g e

The general form of logit regression model is as follow: P (y = 1) = α + β 1 x 1 + β 2 x 2 +...+ β k x k Β i are the coefficient of the model and x i is the characteristics of firm i Therefore, in order to test the first and second hypotheses, this paper uses the following logit regression model: P (y = 1) = α + β 1 ST exp + β 2 Fsales + β 3 Net investment risk + β 4 FD risk + Control + Ƹ The main independent variables are the various measures of foreign currency exposure. STexp is the short term transaction dummy that equals to 1 if company s operations generate any nonfunctional currency transaction, such as vendor payments, sales of goods and services and any repatriation of income profits, dividends, fees, royalties and interest from foreign subsidiaries that eventually raises the firm s short term exposure. Foreign sales by origin ratio (Fsales) are used as a proxy for exposure arises from possession of foreign asset (Geczy et al., 1997). Further, this paper also uses indicator variable that equals to 1 (Net investment risk) if a firm indicates any of its major foreign subsidiaries have non US dollar-functional currency. Changes in the functional currency of major foreign subsidiaries raises the volatility in stockholders equity of a firm, therefore net investment in foreign subsidiary is also a proxy for currency risk. Lastly, Fdebtrisk is the foreign borrowing dummy that equals to 1 if firm indicates any existence of foreign denominated debt for purpose other than hedging activity. This variable is also a proxy for currency risk as firm is able to raise capital by issuing debt in foreign currency. Consistent with the basis risk theory describe earlier, the currency exposure variables is expected to be positive and significantly associated with any category of financial hedging instrument dummy. 24 P a g e

4.2.2 Relationship between financial hedging instruments In order to determine the relationship between each financial hedging instruments, this paper also use the multinomial logit regression in order to test the probability of firm choose one of the categories of financial hedging instruments given its exposure type. This multinominal logit regression is use to confirm the result from the earlier logit regression for the second hypothesis. The multinominal logit regression is similar with Clark & Judge (2009) and Geczy et al. (1997) and can be written as follow: P (y 1 = j) = α + β 1 ST exp + β 2 Fsales + β 3 Net investment risk + β 4 FD risk + Control + Ƹ j=1,2,3,4, for j is the groups of financial hedging method. For the multinomila logit regression, this paper identifies four groups of financial hedging method. Group 1 is those firms that hedge currency risk by using short term foreign currency derivatives (SHFCD) only that include the forwards or options users. Group 2 is those firms hedge that hedge currency risk by using cross currency swap only or swap in combination with other instrument. Group 3 is those firm that hedge currency risk by using foreign currency debt (FD) only (exclude those that used FCD). Group 4 is those firms that hedge currency risk by using both FCD and SHFD. All financial hedging groups are normalized to non-hedgers (Group 0). Similar to the previous logit regression, the main independent variables are the various measures of foreign currency exposure. Each currency exposure variables is expected to be positive and significantly associated with a particular group of financial hedging methods, to provide supporting evidence for the complementary hypothesis. 4.2.3 Financial hedging and firm value In this section this paper investigates the impact of financial hedging on firm value, measured by Tobin s q. This paper undertakes similar basic model with as Allayannis et al. (2001), Allayannis et al. (2011), Allayannis & Weston (2001), Carther et al. (2006) and Kim et al. (2006) that regress the firm value with financial hedging dummy and control variables that also give impact to firm value. The basic OLS regression model is as follow: Tobin s g = α 0 + β 1 Finhedge + β 2 ST exp + β 3 Fsales + β 4 Net investment risk + β 5 FD risk + Control + Ƹ 25 P a g e