Asymmetric Foreign Exchange Exposure and Foreign Currency Denominated Debt: International Evidence

Transcription

1 Asymmetric Foreign Exchange Exposure and Foreign Currency Denominated Debt: International Evidence Sung C. Bae* and Taek Ho Kwon This version: May 2010 * Corresponding author; Tel) ; ) bae@bgsu.edu Bae is the Ashel G. Bryan/Huntington Bank Professor at the Department of Finance, College of Business Administration, Bowling Green State University in Bowling Green, OH. Kwon is a Professor at the Department of Business Administration, Chungnam National University in Daejon, Korea. We thank Dae Woo Cho, Jae Man Chung, Chang Soo Kim, Dan Klein, Mingsheng Li, and Hyung Rae Park for many helpful comments. The work was supported by the Korea Research Foundation Grant funded by the Korean Government (KRF A-B00083). The usual disclaimer applies.

2 Asymmetric Foreign Exchange Exposure and Foreign Currency Denominated Debt: International Evidence Abstract We examine measurement and determinants of asymmetric foreign exchange exposure with a focus on the role of firms usage of foreign currency denominated debt. Employing a large sample of Korean manufacturing firms during , we find significant asymmetries in foreign exchange exposure for Korean firms. We also find that Korean firms with dollar-denominated debt exhibit substantially lower asymmetries in foreign exchange exposure than Korean firms without such debt. Consistent with our hypothesis, our results show that Korean firms dollar-denominated debt plays as a key determinant of their asymmetric foreign exchange exposure. Most interestingly, the regression results show that both a firm s export ratio and dollar-denominated debt ratio are significantly related to the firm s asymmetric foreign exchange exposure but in the opposite direction. These results provide strong evidence that increased asymmetries in foreign exchange exposure resulting from exporting activities can be effectively reduced by the usage of dollar-denominated debt. Our results offer a broadly-applicable implication that firms with high asymmetric foreign exchange exposure can effectively manage their foreign exchange risk from operating activities by selectively using exporting and foreign currency-denominated debt financing. JEL Classification: F31; G15 Key words: Asymmetric foreign exchange exposure; Foreign currency denominated debt; Korean manufacturing firms

3 1. Introduction A large body of studies has examined foreign exchange risk with mixed evidence on the existence of foreign exchange exposure (e.g., Jorion, 1990; Bodner and Gentry, 1993; Chow et al., 1997a, 1997b; Bae et al., 2008). The lack of conclusive empirical evidence on significant foreign exchange exposure may be attributed to the insensitivity of firms operations to foreign exchange rate changes, the firms effective management of their foreign exchange risk, or the problems associated with the measurement of foreign exchange exposure. Bartov and Bodnar (1994) note three main reasons for failing to show a clear relation between foreign exchange rate changes and firm value: (1) the difficulty in identifying foreign exchange rate changes as temporary or permanent phenomenon; (2) the difficulty in collecting data necessary to investigate this relation; and (3) the asymmetric nature of foreign exchange exposure. Carter, Pantzalis, and Simkins (2003) note inappropriate specifications of symmetric regression models of foreign exchange exposure used in previous studies. Similarly, Muller and Verschoor (2006) indicate that previous weak evidence may be attributed at least in part to the failure to account for the asymmetric nature of foreign exchange exposure. In this paper, we examine the measurement and determinants of asymmetric foreign exchange exposure. Under the circumstances of asymmetric foreign exchange exposure where firm value is exposed asymmetrically to the same degree of an increase and decrease in foreign exchange rates, a firm s symmetric benefit-loss structure would not allow the firm to effectively mange its foreign exchange rate risk. This is because the hedging tools that firms use for managing foreign exchange risk in general have symmetric benefit-loss structures except for option contracts. Similarly, the asymmetric risk and hence asymmetric benefit-loss structures of investment assets should be an important factor to consider when investors develop their investment strategies including risk management. 1 In spite of the practical importance of recognizing asymmetric foreign exchange exposure, there has been inadequate research done in identifying and measuring asymmetric foreign exchange exposure and 1 For example, investors would need to assess the performance of investment assets in two different scenarios of increase and decrease in exchange rates, rather than merely at the average rate of exchange rate changes. 2

4 analyzing its determinants. The existing literature shows several factors such as firms pricing policies (mark-up adjustment or pricing-to-market), asymmetric hedging strategies, and possession of real options as potential causes of asymmetric foreign exchange exposure (see, e.g., Miller and Reuer, 1998a; Carter et al., 2003; Koutmos and Martine, 2003). 2 In particular, several studies attempt to explain the asymmetric foreign exchange exposure from the perspective of firms exporting behavior (Allayannis and Ihrig, 2001; Bodner, et al., 2002; Campa and Goldberg, 1994). These studies offer some evidence that firms markup-adjustments and pricing-to-market contribute to the asymmetries in foreign exchange exposure. This evidence is to some extent consistent with firms behavior of increasing mark-ups of exporting goods when the foreign exchange rate increases, allowing exporting firms to gain additional asymmetric benefits. In this paper, we argue that firms usage of foreign currency denominated debt also contributes to the asymmetric foreign exchange exposure but in the opposite way to the firms pricing-to-market policies. While firms pricing-to-market policies magnify the positive aspect of foreign exchange exposure, firms usage of foreign currency denominated debt magnifies the negative aspect of exposure. To investigate the role of foreign currency denominated debt on asymmetric foreign exchange exposure, we develop and test the foreign currency-denominated debt (FCDD) effect hypothesis. The FCDD effect hypothesis is drawn from the leverage effect hypothesis which is widely referred to in the finance literature to explain the asymmetries in stock return responses to new market information (Black, 1976). The basic proposition of the leverage effect hypothesis is that investors responses to good and bad news are not symmetric in terms of the changes in stock prices and risk. Similarly, the FCDD effect hypothesis posits that firm values (and stock returns) change asymmetrically (in different directions) in response to increases and decreases in foreign exchange rates due to the different leverage effect associated with foreign currency denominated debt. The foreign exchange rate changes (increases and decreases) in the FCDD effect hypothesis play roles equivalent to 3 2 Koutmos and Martin (2003) include hysteresis as additional factor in explaining the asymmetric foreign exchange exposure in their study, which is, however, closely related to the pricing-to-market factor. 3 According to the leverage effect hypothesis, investors demand much greater risk premiums (in absolute value) due to price declines and risk increases in response to bad news while they accept much less benefits from price increases and risk declines in response to good news. 3

5 the information (good and bad news) that affects stock returns in the leverage effect hypothesis. According to the FCDD effect hypothesis, the degree of a decrease in stock return variance associated with increases in foreign exchange rates (e.g., won/usd) is greater in absolute value than the degree of an increase in stock return variance resulting from decreases in foreign exchange rates. We focus on manufacturing companies in Korea, one of the premier developing countries. Since Korean manufacturing firms have long depended on international trades and foreign capital, and thus their firm values are highly sensitive to foreign exchange rate changes, Korean manufacturing firms offer an invaluable opportunity to investigate the existence and determinants of asymmetric foreign exchange exposure. Since the Asian financial crisis in late 1997, Korean firms have been exposed to an unprecedented level of foreign exchange risk primarily due to the adoption of the flexible foreign exchange rate system on December 16, 1997 and the increase in their global business operations including outward foreign direct investments. For example, the rate of Korean currency (won) relative to the US dollar has swung between a low of W on April 29, 2005 and a high of W1, on January 9, During our study period of January 1, 1998 to December 31, 2005, the exchange rate started at W and ended at W1, In addition, due to the adoption of the new accounting system that requires Korean firms to report gain and loss in asset values associated with foreign exchange rate changes in the current year s balance sheet, the financial performance of Korean firms is affected to a greater extent by the changes in exchange rates in the same year. Furthermore, the increased role of foreign investors in the Korean capital markets and the availability of foreign stocks and mutual funds to Korean investors suggest that foreign exchange risk is a crucial factor to consider in designing investors investment strategies for global portfolios. 4 Employing 337 sample firms during , we find that there exists significant asymmetric foreign exchange exposure for Korean manufacturing firms. We also find that Korean firms using dollar- 4 Kwon et al. (2005) show that foreign investors of ADRs price foreign exchange risk into the underlying firm s stock returns differently than local investors, indicating the importance of considering firms foreign exchange risk exposure in assessing firm value. 4

6 denominated debt exhibit substantially lower degrees of asymmetries in foreign exchange exposure than Korean firms using no dollar-denominated debt. Most interestingly, our regression results show that a firm s export ratio and dollar-denominated debt ratio are significantly related to the firm s asymmetric foreign exchange exposure but in the opposite direction. As expected, a firm s export ratio has a positive regression coefficient, indicating that firms with higher export ratios tend to have greater asymmetric foreign exchange exposure. This result is consistent with the evidence from previous studies (see e.g. Allayannis and Ihrig, 2001; Bodner, et al., 2002; Campa and Goldberg, 1994) and is largely attributed to the exporting firms behavior of raising mark-ups of exporting goods when the foreign exchange rate increases, resulting in additional asymmetric benefits for exporting firms. In contrast, a firm s dollar-denominated debt ratio carries a negative regression coefficient. Hence, firms with a larger proportion of dollar-denominated debt are likely to have lower asymmetric foreign exchange exposure. The combined evidence on a firm s export ratio and dollar-denominated debt ratio in the opposite direction strongly suggests that the asymmetric foreign exchange exposure resulting from a firm s exporting activities can be effectively controlled and reduced by the exporting firm s usage of dollar-denominated debt. Hence, firms with high degrees of asymmetric foreign exchange exposure can effectively manage their foreign exchange exposure from operating activities by using exporting and dollar-denominated debt financing selectively. In the next section, we offer literature review and develop our hypothesis. Section 3 presents the measurement of asymmetric foreign exchange exposure. Section 4 reports results on the determinants of asymmetric foreign exchange exposure, with summary and conclusions in Section Related Studies and Development of Hypothesis The research on the asymmetric foreign exchange exposure started with efforts to unveil why no significant foreign exchange exposure for multinational firms is detected in previous studies. Miller and Reuer (1998a) show that a small percentage of U.S. manufacturing firms with significant foreign 5

7 exchange exposure exhibit asymmetric exposure profiles, whose evidence is consistent with the predictions from the option pricing theory and pricing-to-market theory. Iorio and Faff (2000) report similar evidence of asymmetric foreign exchange risk in nine out of twenty-four Australian industries when daily stock returns are examined. Koutmos and Martin (2003) show that about 40% of firms in Germany, Japan, the U.K., and the U.S. exhibiting significant foreign exchange exposure have asymmetric exposure profiles. Carter et al. (2003) find that a substantial number of U.S. multinational companies are exposed to strong and weak dollar states differently. Muller and Vershoor (2008) show that about 16% of U.S. multinational firms examined experience significant positive or negative exposure effects during 1970 to Muller and Verschoor (2006) find that 29% of sample U.S. multinational firms with real operations in foreign countries have significant asymmetric foreign exchange exposure to changes in Latin American exchange rates between 1990 and They further demonstrate that these asymmetries are more pronounced toward large versus small foreign exchange movements than over increasing versus decreasing cycles. While several studies attempt to explain the asymmetries in foreign exchange exposure by firms exporting behavior, the literature has not provided conclusive evidence on the relationship between foreign exchange exposure and pricing-to-market, which firms frequently use as pricing policy. 5 Moreover, even if firms accept the relation between the two, this would not necessarily indicate that the pricing-to-market policy is the main cause of firms asymmetric foreign exchange exposure. Furthermore, following the Korean financial crisis in late 1997, the negative relationship between changes in foreign exchange rate (e.g., Korean won per USD) and firm value has been more pronounced than before. This evidence is not easily explained by Korean firms exporting activities. Considering that foreign exchange exposure is a sort of volatility ratio, it is reasonable to conjecture that the asymmetry of foreign exchange exposure is closely related to the asymmetry of volatility. There has been, however, little effort to explain the asymmetry of foreign exchange exposure from this point of 5 This relationship has been examined in Campa and Goldberg (1999), Allayannis and Ihrig (2001), Bodnar, Dumas and Marston (2002). 6

8 view. The foreign exchange exposure is measured as the covariance between foreign exchange rate changes and stock returns divided by variance of foreign exchange rate changes (Adler and Dumas, 1984). Hence, based on the direction of foreign exchange rate changes, the asymmetry of foreign exchange exposure may be caused by either asymmetric covariance between foreign exchange rate changes and stock returns or asymmetric variance of foreign exchange rate changes or the combined effect of both. From this point of view, we can borrow the empirical evidence on the asymmetry in the volatility of stock returns documented in the existing literature for our analysis in explaining the asymmetry in foreign exchange exposure. Among the hypotheses advanced in the current literature in explaining the asymmetry of stock return volatility, the leverage effect hypothesis posits that the magnitudes of responses of stock prices and stock return volatility to good versus bad information are not symmetric. When negative (positive) information reaches the market, a firm s stock price goes down (up), which in turn worsens (improves) the firm s debt ratio, increases (decreases) shareholders riskiness, and finally increases (decreases) the expected volatility of stock returns (Black, 1976). In this process, according to the leverage effect hypothesis, price-decreasing information causes a much larger increase in stock return volatility, whereas priceincreasing information induces a relatively small increase in stock return volatility. Wu (2001) tests the leverage effect hypothesis along with the volatility feedback effect hypothesis and shows that both hypotheses are useful in explaining the asymmetric volatility of stock returns. Firms often raise capital through foreign currency-denominated debt in order to hedge their foreign exchange exposure, take advantage of lower cost than domestic debt, and/or benefit from expected foreign exchange rate changes (Keloharju and Niskanen, 2001). From the viewpoint of foreign exchange risk, the outcome of raising capital to take advantage of lower cost is also closely related to the speculative nature since the outcome of the transaction will be determined by future changes in foreign exchange rates. Hence, the purpose of raising capital through foreign currency-denominated debt can be for hedging foreign exchange exposure and/or for speculation. Keloharju and Niskanen (2001), Elliot et al. (2003), and Kedia and Mozumdar (2003) provide empirical evidence supporting the notion that raising 7

9 capital through the foreign currency denominated debt decreases the foreign exchange exposure. In a study of manufacturing firms in Finland, however, Keloharju and Niskanen (2001) show that firms issue foreign currency denominated debt for both hedging and speculative purposes. Nguyen and Faff (2006) offer similar evidence for Australian firms. The leverage effect hypothesis that explains asymmetric volatility in stock returns offers a theoretical foundation in explaining the role of FCDD in the firms asymmetric exposure to the changes in foreign exchange rates. Drawing from the leverage effect hypothesis, we develop the foreign currencydenominated debt (FCDD) effect hypothesis, which posits that firm values (and stock returns) change asymmetrically (in different direction) in response to increases and decreases in foreign exchange rates due to the different leverage effect associated with foreign currency denominated debt. On the one hand, if an increase in foreign exchange rates raises the Korean won amount of foreign currency denominated debt and thus lowers the firm value, the foreign currency appreciation is equivalent to the negative information that causes a decrease in firm value as in the leverage effect hypothesis. This will worsen the firm s leverage ratio to a greater extent, and hence magnify the negative correlation between changes in the foreign exchange rates and stock returns. In this case, the increase in expected stock return volatility is likely to yield a large covariance between changes in the foreign exchange rates and stock returns. On the other hand, a decline in foreign exchange rates will decrease a firm s debt amount in local currency and thus corresponds to the positive information that boosts the stock price as in the leverage effect hypothesis. This will improve the firm s debt ratio and reduce the risk to the firm s stockholders. As a result, this will induce a smaller increase in the expected stock return volatility, which in turn yields a smaller covariance between changes in the foreign exchange rate and stock returns. 3. Measuring Asymmetric Foreign Exchange Exposure In this section, we first identify the existence of asymmetries in foreign exchange exposure for Korean manufacturing firms. We then examine potential determinants of asymmetric foreign exchange exposure in the next section. 8

10 3.1. Regression models for the measurement of foreign exchange exposure We estimate foreign exchange rate exposure of firm value using the following time-series regression model for each firm by adding a dummy variable representing the direction of the change in foreign exchange rates to the Jorion s (1990) estimation model (Koutmos and Martin, 2003): R i, t α o, i + β M, irm, t + β f, ifxrt + γ i ( Dt FXRt ) + ε i, t = (1) In equation (1), R i is real stock return for individual firm i; R M is real market return proxied by Korea Composite Stock Price Index (KOSPI); FXR is a change in a real (adjusted by the consumer price index) foreign exchange rate; D is a dummy variable with a value of 1 if the foreign exchange rate increases (that is, FXR > 0), and 0 if the rate declines (that is, FXR < 0); and ε i,t is i.i.d. (independently and identically distributed) error term. The key variable in equation (1) is γ, the coefficient that represents the degree of asymmetry in foreign exchange exposure associated with changes in foreign exchange rates in different directions. A significant γ implies that the foreign exchange exposure associated with an increase in the real foreign exchange rate is not symmetric to that associated with a decrease in the real foreign exchange rate. Since the dummy D has a value of 1 when the foreign exchange rate increases and 0 when the rate declines, the foreign exchange exposure associated with a decrease and an increase in the foreign exchange rate is measured by βf and β f + γ, respectively. Considering that the US dollar is the primary foreign currency that Korean firms deal with in their overseas business activities, we compare the estimation of equation (1) between firms that have dollar-denominated debt and firms that have no such debt Data and measurement of variables The sample of our study includes all manufacturing companies listed on the Korea Stock Exchange (KSE) during the period of January 1998 to December Accordingly, we exclude financial companies. Our sample period focuses on the post period of Korean financial crisis in late 1997 in order 9

11 to avoid any compounding effects of the crisis on firm value. 6 With all the necessary information, the final sample consists of a total of 337 manufacturing firms in Korea, spanning sixteen manufacturing industries. In equation (1), the foreign exchange rate is expressed as Korean won per US dollar (won/usd) and thus FXR, change in a real foreign exchange rate, is measured by changes in the monthly average exchange rates of daily exchange rates adjusted by the difference in inflation rates (proxied by consumer price index) between Korea and the U.S. Hence, the change in real won/usd is computed by: change in nominal won/usd (inflation rate in Korea inflation rate in the U.S.). Ri, real stock return of firm i, includes dividends paid using the data from the Korea Securities Research Institute. R M, real market return, is proxied by returns on the Korea Composite Stock Price Index (KOSPI). Both R i and R M are adjusted by the risk-free rates of return, which are proxied by the interest rate on 1-year monetary stabilization bonds (MSBs) issued by the Korean government. 7 The data on foreign exchange rates and MSBs are obtained from the database of the Bank of Korea Distribution of regression coefficients of foreign exchange exposures Table 1 reports the distributions of the regression coefficients (β f,i and γ i ) of foreign exchange exposure estimated from equation (1). Panel A shows the results for two groups of Korean manufacturing firms: firms with dollar-denominated debt and firms without such debt. The first column lists eight categories classified by the magnitude of estimated foreign exchange exposure coefficients. The second and third columns present the number of firms whose coefficient of foreign exchange exposure (β f,i ) and coefficient of asymmetric foreign exchange exposure (γ i ), respectively, lie in the range as given in the first column for firms with dollar-denominated debt. The fourth and fifth columns show the same statistics for firms without dollar-denominated debt. For example, the figures of 6(4) in the second row 6 Since the Korean financial crisis in late 1997, there have been significant changes and reforms across the financial, corporate and public sectors. Furthermore, the Korean currency markets experienced significant volatilities along with a transformation to the flexible exchange rate system during the Korean financial crisis. 7 The monetary stabilization bond is a bond issued by the Bank of Korea, the central bank in Korea, to adjust the money supply and stabilize the interest rate. 10

12 of the second column indicate that six out of 200 firms with dollar-denominated debt have their estimated coefficients of foreign exchange exposure between -3 and -2, and four of the six firms have statistically significant (at least at the 10% level) coefficients. Looking at the mean value of foreign exchange exposure coefficients, Korean firms with dollardenominated debt have on average relatively small coefficients of FX exposure of and asymmetric foreign exchange exposure of in response to a decrease in won/usd rate (that is, an appreciation of the value of won relative to USD). In contrast, Korean firms without dollar-denominated debt have on average a negative and much larger (in terms of absolute value) coefficient of foreign exchange exposure of , indicating that the value of a Korean firm with no dollar-denominated debt increases to a great extent in response to a decrease in won/usd rate. Similarly, Korean firms without dollar-denominated debt exhibit a much larger coefficient of foreign exchange exposure of (= ) than Korean firms with dollar-denominated debt do (0.290 = ) in response to an increase in won/usd rate. The evidence on the negative foreign exchange exposure coefficient of is at least in part due to the fact that the sample firms with no dollar-denominated debt also include firms with dollar-denominated asset only. To examine this conjecture, we compare the coefficients of foreign exchange exposure between firms with dollar-denominated debt and firms with dollar-denominated assets. As shown in Panel B, firms with dollar-denominated asset has an average coefficient of , whereas firms with dollar-denominated debt has an average coefficient of in response to a decrease in won/usd rate. These results offer interesting evidence on Korean firms foreign exchange exposure. Firms without dollar-denominated debt have asymmetric structures of foreign exchange exposure in which their firm values increase with a decrease in won/usd rates but increase to a larger extent with an increase in won/usd rates than firms with dollar-denominated debt. On the contrary, firms with dollar-denominated debt experience a small decrease in firm value with a decrease in foreign exchange rates and a small increase in firm value with an increase in foreign exchange rates. Hence, Korean firms using dollardenominated debt exhibit lower degrees of asymmetry in foreign exchange exposure than Korean firms 11

13 using no dollar-denominated debt. As the foreign exchange rate decreases, so does the value of dollardenominated debt, thus reducing the amount of dollar-denominated liabilities to Korean firms. The results in Table 1 in general support the foreign currency-denominated debt (FCDD) effect hypothesis that Korean firms dollar-denominated debt has the effect of reducing the asymmetric foreign exchange exposure of Korean firms. These results suggest that Korean manufacturing firms can effectively manage their asymmetric foreign exchange exposure caused by asymmetric mark-up adjustments in exporting activities through financing dollar-denominated debt. 4. Analysis of Determinants of Asymmetric Foreign Exchange Exposure As we find preliminary evidence that Korean manufacturing firms with dollar-denominated debt exhibit substantially lower degrees of asymmetry in foreign exchange exposure than Korean firms without such debt, we now examine the role of dollar-denominated debt in a more rigorous regression analysis. Since the degree of the asymmetry in a firm s foreign exchange exposure is also likely affected by several factors other than foreign currency denominated debt, we include variables representing firm characteristics such as export ratio, firm size, foreign direct investment ratio, R&D ratio, hedging activities using derivatives, and industry classification. In this analysis, we test the empirical relationships between a firm s asymmetric foreign exchange exposure and these variables to uncover potential determinants of asymmetric foreign exchange exposure of Korean manufacturing firms Regression model and measurement of variables In order to examine the effect of foreign currency-denominated debt and other firm characteristics on the asymmetric foreign exchange exposure of Korean firms, we estimate the following multivariate regression model: 15 i = δ o, i + δ1dddri + δ 2EXPTi + δ 3FSIZEi + δ 4FDIi + δ 5RNDi + δ 6OPTIONi + k= 1 ˆ γ λ IND + ε (2) k i, k i 12

14 where γˆ i is the coefficients of asymmetric foreign exchange exposure of firm i estimated from equation (1); DDDR is dollar-denominated debt ratio; EXPT is export ratio (relative to total sales); FSIZE is firm size, measured as natural log of firm value; FDI is foreign direct investment ratio (relative to total assets); RND is R&D expense ratio (relative to total sales); OPTION is option trading index; IND is industry dummies, spanning sixteen Korean industries. Below we present the measurement of these variables used in the regression model. DDDR (Dollar-denominated debt ratio) Since the US dollar is the primary foreign currency that Korean firms deal with in their overseas business activities, we use dollar-denominated debt as representative of foreign currency denominated debt. DDDR is measured as the difference between dollar-denominated liabilities and assets divided by total assets. 8 Following the FCDD effect hypothesis, we expect a negative relationship between DDDR and the asymmetric foreign exchange exposure. EXPT (Export ratio) We measure a firm s export ratio as a proportion of total export amount to total sales. Existing studies attempt to explain the asymmetry in foreign exchange rate exposure from the perspective of the pricing-to-market (PTM) in the firm s exporting activity. This approach is a natural one in the sense that the studies of pass-through which is opposite to the PTM have started with an interest in the asymmetric relationship between changes in foreign exchange rates and balance of payments. Mark-up adjustment coefficients are used to measure and analyze the changes in exporting goods PTM according to changes in foreign exchange rates. Several previous studies have attempted to estimate mark-up adjustment coefficients and analyze their determinants (Krugman, 1987; Ohno, 1989; Knetter, 1993). The analyses of the relationship between firms mark-up adjustments and foreign exchange exposure have, however, focused primarily on the theoretical aspect such as model development (Campa and Goldberg, 1999; Allayannis and Ihrig, 2001; Bodnar et al., 2002). These studies are in supportive of the notion that mark- 8 Accordingly, a negative DDDR is possible if a firm has more dollar-denominated assets than liabilities. 13

15 up adjustments are related to foreign exchange exposure under certain assumptions. Little empirical analysis on the relationship between the two variables, however, has been done due to the difficulty in measuring firms mark-up adjustment coefficients. It is expected that if a firm s asymmetric mark-up adjustments cause its asymmetric foreign exchange exposure, an increase in a foreign exchange rate (or a depreciation of Korean won value) will result in an increase in the positive foreign exchange exposure. In other words, if a firm engages in asymmetric mark-up adjustments in which the firm increases the mark-up to a larger extent with an increase in a foreign exchange rate, the firm is expected to experience asymmetric foreign exchange exposure such that the firm s value increases to a greater extent when the foreign exchange rate increases. Accordingly, we expect a positive relationship between a firm s export ratio and its asymmetric foreign exchange exposure. FSIZE (Firm size) We measure firm size as natural log of the market value of common stock. Existing studies on foreign exchange risk show that the size of a firm is closely related to the degree of foreign exchange exposure with contrasting evidence. On the one hand, Chow et al. (1997b) postulate that since a large firm has an advantage in implementing foreign exchange risk management and hedging strategies, a larger firm is associated with smaller foreign exchange exposure. On the other hand, Ware and Winter (1988) predict that a smaller firm has higher financial distress cost and thus a greater incentive for hedging, resulting in smaller foreign exchange exposure. Consistent with Ware and Winter s prediction, He and Ng (1998) find that larger Japanese firms have greater foreign exchange exposure than smaller Japanese firms. Dominguez and Tesar (2005) show similar evidence. In sum, large firms are likely exposed to more risk due to its diverse domestic and overseas business operations but are more effective in managing firm costs and more flexible in dealing with firm risk than small firms. How the size of a firm is related to the foreign exchange exposure is an empirical question. FDI (Foreign direct investments) 14

16 We measure FDI as a proportion of a firm s foreign direct investments to total assets. From the viewpoint of natural hedges in managing foreign exchange exposure, a firm s foreign direct investment activity may b DDDR is measured as the difference between dollar-denominated liabilities and assets divided by total assets. 9 e related to the firm s asymmetric foreign exchange exposure. This is because a firm can secure flexibility and freedom in managing foreign exchange exposure through its foreign direct investments. If a firm has the flexibility in choosing between a foreign production facility and a domestic one, the foreign direct investments would be closely related to asymmetric foreign exchange exposure. For example, if a Korean firm possesses a foreign production facility, exports raw materials to the facility, and imports the finished product back, the firm will cope with the depreciated value of Korean currency by reducing the import of finished products from the foreign facility and increasing domestic production at the same time when the foreign exchange rate (won/foreign currency) increases. In this case, the magnitude of firm value affected by the changes in foreign exchange rates will be different between appreciation and depreciation of Korean won relative to foreign currency. Miller and Reuer (1998b) note the existence of real options along with the PTM as the potential causes of asymmetric foreign exchange exposure. They also note that if a firm s foreign subsidiary possesses foreign currency-denominated debt, it would increase the degree of asymmetry in foreign exchange exposure of the firm in a similar manner to the foreign currency-denominated debt owned by the domestic parent firm. It is an important empirical question whether and how the management of foreign exchange risk using real options such as foreign direct investments is related to the asymmetry of foreign exchange exposure. RND (R&D expenses) We measure RND as a proportion of a firm s R&D expenses to total sales. A firm s R&D expense proxies for the firm s future growth potential. A firm with higher growth potential is more likely to engage in hedging the firm s operating and financing risk using financial derivatives such as options (Lin et al., 2008), reducing asymmetric foreign exchange exposure. Hence, we expect a negative relationship between a firm s R&D expense ratio and asymmetric foreign exchange exposure. 9 Accordingly, a negative DDDR is possible if a firm has more dollar-denominated assets than liabilities. 15

17 OPTION (Option trading index) If a firm engages in hedging activities by utilizing tools that possess an asymmetric benefit-cost structure of foreign exchange exposure, the change in foreign exchange rates would bring in asymmetric influence to firm value. Firms can hedge their foreign exchange exposure using both external market tools including futures and options and natural hedges resulting from the firms internal operating activities. As a proxy for firm s hedging activities, we use a firm s options trading. OPTION represents the frequency of option trading of Korean firms. We measure OPTION by examining the section of transactions (purchases and sales) of financial derivatives in each firm s annual operating reports during the sample period. For example, if a firm reports transactions of foreign currency options for one or two years, a value of 1 or 2, respectively, is assigned to the firm. 10 The relationship between a firm s option trading and asymmetric FX exposure is based on the asymmetric benefit structure of options. Hence, if a firm uses options for hedging purpose, the degree of asymmetry in the firm s foreign exchange exposure will be reduced, resulting in a negative relationship between OPITION and a firm s asymmetric foreign exchange exposure. Based on the above discussions, the expected relationships between each variable and the coefficient of asymmetric foreign exchange exposure along with their measurements are summarized in Table Summary statistics of key variables Table 3 reports summary statistics of three different debt ratios and several key variables of Korean manufacturing firms. Won-denominated debt ratio is computed as the difference between total debt and all foreign-currency denominated debt divided by total assets. All variables represent the annual average of all firms in the sample. Dollar-denominated debt for Korean firms represents the largest foreign currency denominated debt with an average of 2.3% of total assets and a range of -41.2% to 64.0%, dominating all other foreign 10 We also measure OPTION as a dummy variable of 1 if a firm engages in transactions of financial derivatives and 0 otherwise, regardless of the number of years of transactions. The results from the alternative way are qualitatively the same as those reported in our paper. 16

18 currency debt. In contrast, yen-denominated debt represents much less than 1% of total assets of Korean firms. On annual average, a typical Korean firm in our sample has an export ratio of 28.3%, a foreign direct investment ratio of 8.2%, and an R&D ratio of 1.3% Regression results Table 4 presents regression results for the whole sample firms. Due to the possibility of heteroskedasticity problem associated with the cross-sectional regression analysis, t-statistics are heteroskedasticity-adjusted (White, 1980). Looking first at the control variables in the regression, none of the control variables except for EXPT carries a significant coefficient, indicating little impact of firm size, FDI ratio, R&D ratio or option trading on the firm s degree of asymmetric foreign exchange exposure. Among the industry dummies, the regression coefficients for three industries of chemical products, fabricated metal products, and motor vehicles are statistically significant at least at the 10% level. EXPT has a positive and significant (at the 5% level) regression coefficient; hence, firms with higher export ratios tend to have greater asymmetric foreign exchange exposure. This result is consistent with our hypothesized positive relationship of a firm s export ratio to its asymmetric foreign exchange exposure that when the foreign exchange rate (won/usd) increases, exporting firms gain additional benefits by increasing mark-ups of exporting goods. The regression results on EXPT are also in line with those in the previous studies. As a firm has a higher export ratio, the magnitude of increases in firm value associated with increases in foreign exchange rates is greater than that of decreases in firm value associated with decreases in foreign exchange rates, hence increasing the firm s asymmetric foreign exchange exposure. The regression coefficient of the key testing variable DDDR (dollar-denominated debt ratio) is negative and statistically significant at the 1% level, consistent with the hypothesized negative relationship between DDDR and the dependent variable. This result strongly suggests that firms with a larger proportion of dollar-denominated debt relative to total debt are likely to have lower asymmetric foreign exchange exposure. Hence, this result provides evidence that dollar-denominated debt plays an 17

19 effective role in reducing Korean firms asymmetric foreign exchange exposure associated with changes in foreign exchange rates. Combined with the results on EXPT, the results on DDDR suggest that the asymmetric foreign exchange exposure resulting from a firm s exporting activities can be effectively controlled and reduced by the firm s usage of dollar-denominated debt. In particular, firms with high degrees of asymmetric foreign exchange exposure can effectively manage their foreign exchange exposure from operating activities by using exporting and dollar-denominated debt financing selectively Additional regression analysis with different debt ratios In order to gain additional insights into the effects of foreign-currency denominated debt on the firm s asymmetric foreign exchange exposure, we estimate regression equation (2) with two additional debt ratios of (Korean) won-denominated debt ratio and (Japanese) yen-denominated debt ratio, along with dollar-denominated debt ratio. We examine these debt ratios to see whether domestically raised won-denominated debt and yen-denominated debt play a role in explaining Korean firms asymmetric foreign exchange exposure. The regression results are reported in Table 5. In Models 1 and 2, each of the two foreign currency debt ratios (dollar- and yen-denominated debt ratios) enters separately along with won-denominated debt ratio and other control variables. In Model 3, all three debt ratios are used. Unlike dollar-denominated debt ratio, won-denominated debt ratio and yen-denominated debt ratio have negative but insignificant (at the 10% level) regression coefficients in all three regressions. Hence, neither domestically raised wondenominated debt nor yen-denominated debt is related to Korean firms asymmetric foreign exchange exposure. These results suggest that neither of these two types of debt is an effective tool in managing Korean firms asymmetric exposure to changes in foreign exchange rates. In contrast, dollardenominated debt ratio has a negative and significant coefficient in Models 1 and 3, confirming our evidence on the effective role of dollar-denominated debt in managing Korean firms asymmetric nature of foreign exchange exposure. 18

20 4.5. Robustness test In order to check the robustness of our regression results, we perform a robustness test. As shown in Table 1, the distributions of the coefficients of asymmetric foreign exchange exposure estimated by equation (1) shows some extreme values, which may unduly affect the regression results of equation (2). Considering this possibility, we estimate regression equation (2) for firms whose estimated coefficients of asymmetric foreign exchange exposure lie within -3 and +3 by excluding firms with extreme values (greater than +3 or smaller than -3) of foreign exchange exposure coefficients. 11 Table 6 reports the regression results for the reduced sample. While there are slight differences in the signs and significance levels of regression coefficients of control variables, compared to those reported in Table 4, the main results on the effect of DDDR (dollar-denominated debt ratio) on asymmetric foreign exchange exposure remain unchanged. The regression coefficient of DDDR is negative and significant at the 5% level, consistent with the hypothesized relationship between DDDR and asymmetric foreign exchange exposure that firms with higher dollar-denominated debt ratios tend to have lower asymmetric foreign exchange exposure. Among the control variables, EXPT has the same positive and significant (at the 5% level) coefficient as with the whole sample in Table 4, and RND has a negative and significant (at the 10% level) coefficient. Hence, a firm s asymmetric foreign exchange exposure is positively related to its export ratio but negatively to its R&D expense ratio. The overall results in Table 6 confirm that the earlier regression results with the whole sample remain robust. 5. Summary and Conclusions A firm s foreign exchange exposure becomes asymmetric when firm value changes asymmetrically to the same degree of changes in foreign exchange rates. In spite of the practical importance of recognizing asymmetric foreign exchange exposure, little research has been done in identifying 11 Although not reported in our paper, we obtain similar results by excluding firms whose estimated coefficients of asymmetric foreign exchange exposure are greater than +2 or smaller than

21 asymmetric foreign exchange exposure and analyzing its determinants. Our paper examines these issues by focusing on Korean manufacturing firms, whose values are known to be highly sensitive to foreign exchange rate changes due to their dependence on international trades and foreign capital. In order to examine the empirical relationship between foreign currency denominated debt and asymmetric foreign exchange exposure, we develop and test the foreign currency-denominated debt (FCDD) effect hypothesis, drawn from the leverage effect hypothesis. The FCDD effect hypothesis posits that firm values (and stock returns) change asymmetrically (in different direction) in response to increases and decreases in foreign exchange rates due to the different leverage effect associated with foreign currency denominated debt. Employing 337 Korean manufacturing firms during , our results show that there exist significant asymmetries in foreign exchange exposure for Korean firms. Our results also show that Korean firms using dollar-denominated debt exhibit substantially lower asymmetries in foreign exchange exposure than Korean firms using no such debt. Most interesting are our regression results that a firm s export ratio and dollar-denominated debt ratio are significantly related to the firm s asymmetric foreign exchange exposure but in the opposite direction, positively for export ratio and negatively for dollar-denominated debt ratio. These results strongly indicate that the increased asymmetries in foreign exchange exposure resulting from firms exporting activities can be effectively reduced and offset by their usage of dollar-denominated debt. Hence, when firm values are exposed asymmetrically to the changes in foreign exchange rates, firms can effectively manage the asymmetries in foreign exchange risk from operating activities by using exporting and dollar-denominated debt financing selectively. Our regression results are robust to a different sample specification. 20

22 References Adler, M., Dumas, B., Exposure to currency risk: Definition and measurement. Financial Management 13, Allayannis, G., Ihrig J., Exposure and Markups. Review of Financial Studies 14, Bae, S.C., Kwon, T.H., Li, M., 2008, Foreign exchange exposure and risk premium in international investments: Evidence from American Depositary Receipts, Journal of Multinational Financial Management 18, Bartov, E., Bodnar, G.M., Firm valuation, earnings expectations, and the exchange-rate exposure effect. Journal of Finance 49, Black, F., Studies of stock market volatility changes. Proceedings of the American Statistical Association, Business and Economic Section, Bodnar, G.M., Dumas, B., Marston, R.C., Pass-through and exposure. Journal of Finance 57, Bodnar, G.M., Gentry, W.M., Exchange rate exposure and industry characteristics: Evidence from Canada, Japan and the USA, Journal of International Money and Finance 12, Carter, D., Pantzalis, C., Simkins, B.J., Asymmetric exposure foreign exchange risk: Financial and real option hedges implemented by U.S. multinational corporations. Working paper, Oklahoma State University. Campa, J.M., Goldberg, L.S., Investment, pass-through, and exchange rates: A cross-country comparison. International Economic Review 40, Chow, E.H., Lee, W.Y., Solt, M.E., 1997a. The exchange rate risk exposure of asset returns. Journal of Business 70, Chow, E.H., Lee, W.Y., Solt, M.E., 1997b. The economic exposure of U.S. multinational firms. Journal of Financial Research 20, Dominguez, K.M.E., Tesar, L.L., Exchange rate exposure. Journal of International Economics, INEC

23 Elliott, W., Huffman, S., Makar, S., Foreign denominated debt and foreign currency derivatives: Complement or substitutes in hedging foreign currency risk? Journal of Multinational Financial Management 13, He, J., Ng, L.K., The foreign exchange exposure of Japanese multinational corporations. Journal of Finance 53, Iorio, D.I., Faff, R., An analysis of asymmetry in foreign currency exposure of the Australian equities market. Journal of Multinational Financial Management 10, Jorion, P., The exchange rate exposure of U.S. multinationals. Journal of Business 63, Kedia, S., Mozumdar, A., Foreign currency denominated debt: An empirical examination. Journal of Business 76, Keloharju, M., Niskanen, M., Why do firms raise foreign currency denominated debt? Evidence from Finland. European Financial Management 7, Knetter, M.M., International comparisons of pricing-to-market behavior. American Economic Review 83, Koutmos, L., Martin, A.D., Asymmetry exchange rate exposure: Theory and evidence. Journal of International Money and Finance 22, Krugman, P.R., Pricing to markets when the exchange rates change. Real-Financial Linkages Among Open Economics in J. David Richardson and Sven Arndt (eds.), The MIT Press. Cambridge. Kwon, T.H., Bae, S.C., Chung, J.M., Do foreign investors price foreign exchange risk differently? Journal of Financial Research 28, Lin, C., Phillips, R., Smith, S.D., Hedging, financing, and investment decisions: Theory and empirical tests. Journal of Banking and Finance 45, Miller, K.D., Reuer, J.J., 1998a. Asymmetric corporate Exposures to foreign exchange rate changes. Strategic Management Journal 19, Miller, K.D., Reuer, J.J., 1998b. Firms strategy and economic exposure to foreign exchange rate movements. Journal of International Business Studies 29,