Firmwide Risk Management of Foreign Exchange Exposure by U
Firmwide Risk Management of Foreign Exchange Exposure by U.S. Multinational Corporations David A. Cartera, Christos Pantzalisb, and Betty J. Simkinsa a
Department of Finance, College of Business Administration, Oklahoma State University, Stillwater, OK 74078-4011, USA b Department of Finance, College of Business Administration, University of South Florida, Tampa, Florida 33620-5500, USA Version of April 28, 2003 Abstract This paper investigates the impact of firmwide risk management practices on the foreign exchange exposure of 208 U.S. multinational corporations (MNC) over the period 1994 to 1998. Firmwide risk management is referred to here as the coordinated use of both financial hedges, such as currency derivatives, and operational hedges, described by the structure of a firm’s MNC foreign subsidiary network. We find that the use of currency derivatives, particularly forward contracts, is associated with reduced levels of foreign-exchange exposure. Furthermore, MNCs with dispersed operating networks have lower levels of currency exposure. These findings are robust to alternative ways of measuring foreign exposure. Finally, our results strongly support the view that MNCs hedging in a coordinated manner can significantly reduce exposure to currency risk. These results strongly suggest that operational and financial hedges are complementary risk management strategies.
JEL Classification: F30, F31, G15 Key words: Multinational finance, Risk management, Firmwide risk management
* Please direct all correspondence to: Betty J. Simkins, Department of Finance, College of Business Administration, Oklahoma State University, Stillwater, OK 74078-4011, (405) 744-8625 (office), (405) 744-5180 (fax), email: [email protected].
We thank Cesare Robotti for comments on an earlier draft of this paper. A prior version of this paper was presented at the 2000 Financial Management Association meetings.
Firmwide Risk Management of Foreign Exchange Exposure by U.S. Multinational Corporations Abstract This paper investigates the impact of firmwide risk management practices on the foreign exchange exposure of 208 U.S. multinational corporations (MNC) over the period 1994 to 1998. Firmwide risk management is referred to here as the coordinated use of both financial hedges, such as currency derivatives, and operational hedges, described by the structure of a firm’s MNC foreign subsidiary network. We find that the use of currency derivatives, particularly forward contracts, is associated with reduced levels of foreign-exchange exposure. Furthermore, MNCs with dispersed operating networks have lower levels of currency exposure. These findings are robust to alternative ways of measuring foreign exposure. Finally, our results strongly support the view that MNCs hedging in a coordinated manner can significantly reduce exposure to currency risk. These results strongly suggest that operational and financial hedges are complementary risk management strategies.
I.
Introduction The practice of corporate risk management has changed dramatically over the past two
decades. Originally, risk management was implemented on an uncoordinated basis across different units of the firm. The primary focus of these ad hoc risk management programs was to minimize costs of particular units. Today, however, risk management of currency exposure has, in many cases, evolved into a firmwide exercise that addresses both short-term and long-term exposures and encompasses financial as well as operational hedges. The ultimate goal of firmwide risk management is to reduce risk while placing the firm in a position to benefit from opportunities that arise from exchange rate changes.1 For example, Davis and Militello (1995) describe how Union Carbide employs a firmwide perspective in risk management. The company uses a one-year horizon for financial hedges (e.g., foreign-exchange derivatives), whereas for longer horizons, operational adjustments are made in sourcing, utilization of different plant locations, and pricing. We define firmwide risk management for multinational corporations (MNCs) as the combined use of both financial and operational hedges as part of an integrated risk management strategy aiming at reducing exposure to foreign-exchange risk.2 Changes in exchange rates can
1
See Paul-Choudhury, 1996.
2
Firmwide risk management is one of the most widely used out of a group of synonyms that describe a broad and comprehensive view of managing risk across the firm. Other terms used to describe this type of coordinated risk management are enterprise risk management, global risk management, and strategic risk
1
influence MNCs’ current and future expected cash flows and ultimately, stock prices. The direction and the magnitude of changes in the exchange rate on firm value are a function of the firm’s corporate hedging policy and the structure of its foreign currency cash flows. The latter depends on the firm’s competitive position in the industries in which it operates. The former indicates whether the MNC utilizes operational hedges and financial hedges to manage currency exposure. Beginning with the seminal study by Jorion (1990), initial research in this area focused on whether corporations are exposed to foreign exchange risk (see Bodnar and Gentry, 1993, Bartov and Bodnar, 1994, 1995, and Chow, Lee and Solt, 1997a,b). Allayannis and Ofek (2001) and Simkins and Laux (1996) investigate the effect of financial hedging on foreign-exchange exposure. More recently, Pantzalis, Simkins, and Laux (2000) examine the ability of operational hedges to reduce exposure.
However, few studies thus far have examined the combined
influence of financial hedges and operational hedges on foreign exchange exposure.3 Our research fills this gap in the literature by examining the relationship of integrated financial and operational hedging programs (firmwide risk management) and foreign exchange exposure for a sample of U.S.-based MNCs. It is common practice among firms to use a combination of production and marketing strategies across the firm’s different operating units (operational hedges) to manage long term exposure, whereas foreign exchange derivatives (financial hedges) are more often used for managing short term exposure. Long-term operating policy adjustments are costly and difficult to reverse, hence they are most effective when the firm possesses a network of multiple operating units that span many business and geographic areas. We describe the firm’s ability to implement operational hedges by examining the network structure of its foreign operations (i.e., the MNC
management. Firmwide risk management addresses exposures on a global basis, in that it considers all parts of the firm and tries to cope with both short-term and long-term exposures. For example, The Tower Group uses the term enterprise risk management and defines it as “the process of managing the risk faced by an institution on a global, institution-wide basis.” (see Deragon, 2000). As reported by Banham (2000), Freeman Wood, Ford Motor Company’s director of global risk management, states: “We want to understand our total aggregate exposure to financial market risks and nonfinancial market risks. The next phase is measuring these risks in such a way as to permit discussion and comparison. That is followed by a determination of which risks to retain, hedge or transfer. Finally, we’ll determine the optimum risk transfers.”
2
network).4 We analyze financial hedges by utilizing information on currency derivatives’ usage. We consider usage of different foreign currency derivatives contracts such as futures, forwards, swaps, options, forward rate agreements, and other derivative contracts that have a foreign currency as the underlying commodity. We argue that the coordinated use of operational and financial hedges should more effectively reduce exposure to foreign exchange risk, because it addresses the firm’s overall exposure both in the short- and in the long-term. The two important contributions of our research are as follows.
First, the evidence
suggests that the coordinated use of operational and financial hedges can effectively reduce foreign currency exposure. To our knowledge, this is the first study to empirically document this result.
Furthermore, these results are robust to alternative methods of measuring currency
exposure, operating hedges, and financial hedges. Second, our results indicate that multinational corporations are taking a firmwide or “strategic” perspective in their currency risk management strategy, and are thus focusing on hedging overall economic exposure. This paper proceeds as follows: Section II examines prior research on foreign-exchange exposure and the use of operational and financial hedging to manage exposure. Section III discusses the methodology we use to investigate the determinants of foreign-exchange exposure by U.S. multinational firms. Our data sources and sample used are described in Section IV, while Section V reports our empirical results. We conclude the paper in Section VI.
II.
Multinational Corporations, Foreign Exchange Exposure, and Firmwide Risk Management A number of studies have examined the uniqueness of U.S. multinational corporations
(MNCs). For example, Errunza and Senbet (1981, 1984), Kim and Lyn (1986), and Morck and Yeung (1991) investigate the value of international operations. However, as Christophe (1997) points out, these studies generally focus on MNC activities during the 1970s, a period
3
The only two previous studies that examined operational as well as financial hedges are Allayannis et al (2001) and Allayannis and Weston (2002). Their findings are discussed in the next section. 4 An MNC’s network has to do with the number of different foreign nations in which operations are conducted (“breadth” of MNC network), as well as the concentration of operations in various locations (“depth” of MNC network).
3
characterized by relatively stable exchange rates. In fact, the exchange rate volatility that has existed since the 1970s makes it necessary to investigate the means by which MNCs manage their risk, and the effectiveness of their risk-management techniques. Financial and operational hedges are two fundamental ways MNCs can manage the risk introduced by increased exchange rate volatility. Flood and Lessard (1986) identify two types of exposure to foreign exchange risk: 1) transaction exposure and 2) operating exposure. Transaction exposure is the effect of unanticipated changes in real exchange rates on nominal cash flows (i.e., cash flows fixed in nominal terms) and primarily a short-term exposure that can be hedged using financial derivatives. In contrast, operating exposure is the effect of unanticipated changes in exchange rates on the cash flows associated with a firm’s real assets and liabilities and is, therefore, primarily a long-term exposure that amounts to the impact of unexpected changes in the exchange rate on the firm’s competitive position. Logue (1995) and Chowdhry and Howe (1999) argue that operating exposure cannot be effectively managed using financial hedges. Instead, they suggest that long-term strategy adjustments (i.e., operational hedges) are the most effective way of managing long-run operating exposure. A firm facing future, contractually fixed, foreign-currency cash flows, in which the only source of uncertainty is the exchange rate (i.e., transaction exposure), can easily hedge with swaps or forward contracts.5 However, if the future cash flows are uncertain and not perfectly correlated with the exchange rate (i.e., operating exposure), financial hedging is likely to be ineffective. A firm’s operating exposure to currency risk depends on the effect of unexpected changes in the exchange rate on the firm’s output prices (e.g., product prices) and input costs (e.g., raw materials, labor costs, etc.). Since the correlation of prices with exchange rates is determined by the degree of segmentation of their respective markets, operating exposure depends on whether input costs and output prices are determined locally or globally. Firms
5 In fact, annual surveys conducted by the Wharton School and Chase Manhattan Bank find that many firms use currency derivatives to manage foreign-exchange risk (see Bodnar, Hayt, and Marston, 1998). Surveys of corporate treasurers also reflect a trend toward management, not only of transaction exposure, but also translation exposure and economic exposure broadly construed (Jesswein, Kwok, and Folks, 1995).
4
facing substantial operating exposures, as is often the case for U.S. MNCs, can manage this exposure by devising operating strategies that consist of combinations of different marketing initiatives, such as market selection or pricing strategy, and production initiatives, such as raw materials sourcing and production location (see Shapiro, 1996). Thus, operational hedges entail long-term operating policy adjustments that are implemented within a firm’s network of operating units. Theoretically, the effectiveness of these policies in managing operating exposures is enhanced when they can be implemented across different lines of business and locations.6 Obviously, not every firm will be able to manage currency risk exposure with the same degree of effectiveness. We argue that firms with operations spread over many currency and business areas are more insulated from foreign exchange exposure, since they have at their disposal a larger set of alternatives to devise effective operational hedges. In effect, because of their geographic diversification, MNCs own a portfolio of real options not available to purely domestic firms. This operating flexibility allows MNCs to effectively react to exchange rate changes by deciding where to shift production, source for inputs, or declare profits among the different locations in which they operate.7 The degree of operating flexibility of a MNC is clearly a function of the structure of its foreign operations’ network. We postulate that MNCs that have operations in a large number of geographic regions will be in a better position to effectively construct operational hedges than similar-sized MNCs whose operations are concentrated in few geographic regions. Most recent empirical studies in the risk management area have examined financial hedging by nonfinancial corporations, including MNCs.
For instance, Nance, Smith, and
Smithson (1993), Dolde (1995), and Mian (1996) examine the hedging policies of corporations. Further, Géczy, Minton, and Schrand (1997) investigate the use of currency derivatives by firms. They find that corporations with extensive exposure to foreign exchange risk are more likely to 6
For example, Cummins Engine Company states they have substantially reduced exposure by building manufacturing plants in three major economic regions of the world – the U.S., Europe, and Japan. National Semiconductor says they take into account economic exposure when selecting manufacturing plant locations (see Davis and Militello, 1995). 7
For a discussion of the MNCs’ operating flexibility see Kogut (1983), Kogut and Kulatilaka (1995), and Buckley and Casson (1998) among others. For empirical evidence regarding the value of operating flexibility see Allen and Pantzalis (1996).
5
use currency derivatives. While these studies focus on the determinants of a firm’s hedging policies, they do not investigate the effectiveness of the policies. To date, only a few studies have investigated the effect of financial hedging on foreign-exchange exposure (see, for example, Allayannis and Ofek, 2001, and Simkins and Laux, 1997). Further, recently there has been more interest in the use of operational hedges by MNCs seeking to reduce exposure (see Pantzalis, Simkins, and Laux, 2001). Most importantly, though, is the fact that there has been very limited empirical evidence on the combined use of financial and operational hedges in foreign exchange risk management. The only study that has provided some preliminary evidence is by Allayannis et al (2001). In their study, Allayannis et al utilize measures of the MNC network (similar to the ones found in Pantzalis et al (2001)) as proxies for the firms’ ability to implement operational hedges. They find that operational hedges are not effective for risk management. Furthermore, they find that the likelihood of financial derivatives’ use increases with the geographic dispersion of the U.S. MNCs’ operations and that operational strategies benefit shareholders when used in combination with financial strategies. However, they restrict their tests to firms with positive currency exposure coefficients and do not control for size and several other factors that may affect exposure. Our research fills this gap in research by examining firmwide risk management of MNCs. Obviously, managing risk across a global firm is a complex task, not only because exposures may be difficult to quantify but also because hedging a group of exposures typically requires a more sophisticated approach than hedging individual exposures. With the development of sophisticated risk management techniques and digital data transfer technology, companies are now able to effectively evaluate and manage firmwide cashflows on a global basis, by incorporating financial and operational hedges into an integrated analysis (Lehtinen, 1996). Proponents of a firmwide approach to risk management argue that potential benefits from it can be profound.
By viewing decisions related to financial transactions and operations as
interrelated, risk can be reduced, controlled or transferred to minimize variations that affect performance.
6
III.
Methodology
A.
Measuring Foreign Exchange Exposure We use a two-factor market model to estimate the effect of exchange rate changes on
the stock returns for our sample firms. This methodology is standard in other research examining currency exposure (e.g., Jorion, 1990, Bodnar and Gentry, 1993, Bartov and Bodnar, 1994, Choi and Prasad, 1995, Simkins and Laux, 1997, and Bodnar and Wong, 2003). For each firm in our sample, we estimate the following time-series regression: Rit = αi + βiRmt + γiRct + εit,
(1)
where Rit is the rate of return on the ith company’s common stock in month t, Rmt is the return on the market index, Rct is the return on the exchange rate index, and εit is the idiosyncratic error term.
The exchange rate is measured as the price of the U.S. dollar in foreign currency;
therefore, a positive value for Rct indicates a strengthening dollar. For each firm, the estimated coefficient γi is a measure of the sensitivity of a firm’s stock returns to changes in foreignexchange rates, and can be either positive or negative. We estimate the stock-price sensitivity to changes in exchange rates for our sample firms for the period 1994-98. This sample period is chosen for two reasons. First, we know which firms actively use currency derivatives over this period based on reporting disclosures and limit our study to a period where hedging practices are most likely to be consistent. Second, prior studies by Bartov and Bodnar (1995), Allayannis (1996), and Chow, Lee, and Solt (1997a) find that a firm’s exposure changes over longer time frames. Among other factors, this is likely due to changes in a firm’s MNC network. To avoid time-varying exposure coefficients, we limit our study to a five-year period. B.
Determinants of Foreign Exchange Exposure
We hypothesize that financial and operational hedges can be effective in reducing foreign exchange exposure. To test this hypothesis, we analyze the following cross-sectional regression: ⎮γ i⎮ = δ0 + δ1SIZEi + δ2LTD/TAi + δ3INSDPi + δ4INSDPSQi + δ5BLOCPi + δ6INSTPi + δ7MNCi + δ8DER/FSi + εi,
7
(2)
where ⎮γi⎮ is the absolute value of the ith firm’s foreign exchange exposure coefficient estimated in Equation (1).8 Since, as discussed earlier, firms can be either negatively or positively exposed to currency risk, we use the absolute value of the estimated exposure coefficient as dependent variable and estimate δ7 and δ8 to measure the effectiveness of operational and financial hedges in reducing the firm’s absolute currency exposure. SIZEi is measured by the firms’ end of year total assets (in millions of dollars). LTD/TAi is long-term debt to total assets. INSDPi is percentage of common stock outstanding held by insiders; INSDPSQi is INSDPi squared. BLOCPi is the percentage of stock held by blockholders. INSTPi is the percentage of stock held by institutions; MNCi is a proxy for operational hedging, and DER/FSi is the total notional value of currency derivatives divided by foreign sales. Table 1 provides a summary of the variables used in this study as well as the expected relationships between these variables and the absolute value of the estimated foreign exchange coefficient. [Insert Table 1 about here.] Size is included in the model as a proxy for the firm’s expertise in devising hedges (see Booth, Smith, and Stolz (1984)). Thus, we expect a negative relationship between size and exposure. Géczy, Minton, and Schrand (1997) provide evidence that firms hedge to reduce the costs of financial distress. We include the long-term debt ratio (LTD/TA) in our model in order to control for the impact of long-term leverage on firms’ exposures. If LTD/TA proxies for expected bankruptcy costs then the relationship between long-term debt ratio and exposure should be negative. This occurs because bondholders may encourage managers of financially distressed firms to hedge in order to reduce cash flow volatility. In addition, LTD/TA may include a foreign-debt component, which provides a natural hedge against foreign currency exposure.9
This would also imply a
negative relationship between LTD/TA and the absolute value of the exposure coefficient. Whidbee and Wohar (1999) document that managerial incentives and external monitoring 8 Jorion (1990) notes that two-step estimation procedures of this type can possibly produce biased standard errors in the second-stage regression. This occurs if errors in (1) are correlated across stocks and exposure coefficients estimated over the same sample period, yielding estimated exposure coefficients that are not independent across equations. However, Jorion (1990) reports that most of the correlation is due to stock market movements, and is purged by the inclusion of the market return in our equation (1).
8
affect the decision to use derivatives. Also, as shown in Doukas, Hall and Lang (1999), if currency risk is priced, then the shareholders’ concern regarding currency-risk exposure, coupled with managers’ risk aversion can explain why firms may elect to implement extensive foreign-exchange risk-management programs.
We include four equity ownership variables (INSDPi, INSDPSQi,
BLOCPi, and INSTPi) in our cross-sectional model to account for the likelihood that the firm’s foreign-exchange exposure may be affected by its ownership structure. Smith and Stulz (1985) point out that managers who are over-invested in their own firms are more likely to hedge risk exposures, because that reduces the uncertainty associated with their own wealth and decreases the likelihood that they are disciplined for poor firm performance. This is supported by evidence from the goldmining industry [Tufano (1996)].
However, Géczy, Minton, and Schrand (1997) fail to find a
significant relationship between currency risk-hedging and managerial incentives. Smith and Stulz (1985) and Froot, Scharfstein, and Stein (1993) argue that value-maximizing firms hedge risk exposures in order to reduce costs associated with agency conflicts between equity- and debtholders (in particular, the under-investment problem). These studies imply that managerial hedging decisions may be value increasing or decreasing, depending on the managerial incentives. The alignment of managerial incentives with those of the shareholders or the lack thereof will affect the hedging behavior of firms. For low managerial ownership levels, we hypothesize that, as managerial ownership increases managers will engage in less excessive hedging. However, when managers own many shares (i.e. when they are entrenched), they would have more incentives to (excessively) hedge in order to reduce their own personal wealth portfolio’s risk (since they are overinvested in their own firm), potentially resulting in less exposure for such firms. Therefore, to capture this non-linear effect, we include both INSDPi, (hypothesize a positive coefficient) and INSDPSQi (hypothesize a negative coefficient) in the regressions.10
9
Kedia and Mozumbar (1999) show that the fraction of debt denominated in foreign currency is strongly related to the degree of foreign exposure. 10 Other studies have used insider-squared variables to account for curvilinear managerial ownership effects. For example, McConnell and Servaes (1990, 1995) find that market value first increases and then decreases with managerial shareholdings. They found that for low levels of managerial ownership there is a convergence of interests between managers and shareholders, while beyond a critical value (usually around 50% ownership –
9
Outsiders, who have the incentive to oversee the manager’s activities because they want to protect their own investment in the firm, can monitor hedging decisions. This is particularly the case for outsiders who own a substantial fraction of the firm’s equity. Blockholders (owners of 5% or more of the common shares outstanding) will monitor managerial hedging decisions and tend to favor risk management activities that reduce exposure. This is because blockholders, like the entrenched managers, are likely to have a disproportionately large fraction of their wealth invested in the firm. Thus, we expect a negative relationship to exist between block-holdings (BLOCPi) and risk-exposure levels. In addition to the effect of blockholders, we consider the impact of institutional shareholders (INSTPi) on exposure.
Although institutional shareholders may not be blockholders, they are
equipped with the resources to analyze managerial actions. However, the ability of institutional shareholders to diversify their portfolios may result in less monitoring intensity with respect to managerial hedging decisions.
Therefore, the expected relationship between INSTPi and the
absolute value of the exposure coefficient is ambiguous. Alternatively, institutional stakes can be viewed as a proxy for short-selling constraints (see, Nagel (2003), among others). Then, stocks of firms with low (high) institutional shareholdings would be harder (easier) to sell short.11 Since demand for shorts increases with stock price level (Figlewski (1981)) and volatility (Boehme et al (2002)) managers may decide to hedge to reduce that demand and the looming subsequent price correction. In other words, if hedging lowers volatility for the stock price and the demand for shortselling, then we would expect firms with lower institutional shareholdings to hedge in order to alleviate the price pressure that arises from the short-sale constraints. We also investigate the effect of different types of currency derivatives on foreign exchange exposure. We accomplish this by replacing the DER/FSi variable that aggregates all types of currency derivatives with four different variables measuring the usage of options, forwards, swaps, and other currency derivatives. This relationship is analyzed in Equation (3) below:
consistent with the model by Stulz (1988)), there is a reversal, due to entrenched managers making decisions that may not be value increasing.
10
⎮γ i⎮ = δ0 + δ1SIZEi + δ2LTD/TAi + δ3INSDPi + δ4INSDPSQi + δ5BLOCPi + δ6INSTPi + δ7MNCi + (3) δ8OPT/FSi + δ9 FWD/FS i + δ10SWAP/FSi + δ11OTH/FSi + εi, where OPT/FSi is measured as the total notional principal amount of currency options divided by foreign sales; FWD/FS i is the total notional principal amount of currency forwards divided by foreign sales; SWAP/FSi is the total notional principal amount of currency swaps divided by foreign sales; and OTH/FSi is the total of other currency derivatives (such as futures, caps, floors, forward rate agreements, etc.) divided by foreign sales. When analyzing both Equations (2) and (3), we employ weighed least squares and weight our independent variables by the reciprocal of the squared standard error of the exposure coefficient from estimating equation (1).
This procedure gives more weight to the exposure
coefficients that are estimated more precisely.
IV.
Data Sources and Sample Selection Data utilized in this study are obtained from a variety of sources. Firm-specific financial
data used in the cross-sectional regressions are obtained from COMPUSTAT for the year 1996, the mid-point of our study. We collect the 1996 common equity ownership data from Disclosure. Data on MNCs’ financial hedging practices are obtained from the SWAPS Monitor database for 1996. We select MNCs from the cross-section of NYSE-, AMEX- and NASDAQ-listed U.S.-based mining and manufacturing firms (i.e., firms with SICs of 3999 or less). A MNC is defined as a firm that has at least one majority owned foreign subsidiary using information collected from Dunn and Bradstreet’s Who Owns Whom. This results in an initial sample of 340 firms. A final sample size of 208 U.S.-based MNCs, is obtained for all MNCs for which complete financial, ownership, and subsidiary location data are available. Security prices are obtained from the University of Chicago CRSP database for the sample period January 1994 to December 1998. Two different measures of the market return, Rmt,, are used: (1) the value-weighted CRSP market portfolio return (VWR) and (2) the equally-weighted CRSP market portfolio return (EWR). The VWR is commonly used in prior studies on exchange
11 Chen, Hong and Stein (2003) show that low breadth of ownership (i.e., low institutional holdings) indicates that few investors have long positions in the stock, which signals that the short-selling constraint is binding, implying high prices relative to the fundamentals and low expected returns.
11
rate exposure, so it is used here for consistency. However, Bodnar and Wong (2003) point out that the VWR can distort the sign and size of the resulting exposures because of an inherent relation between market capitalization and exposure. They state that exposure studies using the VWR control not only remove the “macroeconomic” effects from the exposure estimates, but also cause a distribution shift in the positive direction for exposure estimates.
They show that
replacing the VWR with the EWR in Equation (1) results in residual exposures that are more consistent with the actual cash flow impact of exchange rate changes predicted by corporate finance models. For this reason, Bodnar and Wong recommend using the EWR control to correct for the potential bias of the VWR control and to achieve less distortion in the residual exposures. Changes in international trading relationships makes it necessary to select exchange rate indexes comprised of currencies which most accurately reflect U.S. competitiveness in world markets. Newer indexes created by the Federal Reserve Board (FRB) focus more directly on the dollar’s movements in foreign exchange markets against a broader set of currencies and are designed to reflect the changing structure of trade patterns and exchange rates.
The FRB
updates currency weights in their indexes annually to most accurately measure the U.S. dollar’s competitive value. For these reasons, we select indexes published by the FRB to compute monthly changes in the value of the dollar, Rct.12 Appendix 1 lists these indexes together with the currencies and weights. These indexes are: (1) the nominal Broad Index, which consists of thirty-five currencies of all foreign countries or regions that had a share of U.S. non-oil imports or nonagricultural exports of at least 0.5 percent in 1997, (2) the nominal Major Index, which includes sixteen currencies traded in deep and relatively liquid financial markets and for which short- and long-term interest rates are readily available, (3) the nominal OITP (Other Important Trading Partners) Index, which includes nineteen currencies of key U.S. trading partners in Latin America, Asia, the Middle East, and Eastern Europe, and (4) the real Broad Index, which captures the real (price adjusted) exchange rate value of the Broad Index. The Major and OITP indexes are subsets of the Broad Index.
12 For more information on the currency indexes, refer to Leahy (1998) or go to the Federal Reserve Board website at www.federalreserve.gov.
12
The exchange rate exposure of the sample firms is estimated from Equation (1) using the two different market controls indices (VWR and EWR) with each of the four foreign exchange indexes, resulting in a total of eight different estimates of foreign exchange exposure. This allows us to examine the robustness of the results using the most commonly used market portfolio control (VWR) with results obtained using the EWR market portfolio control recommended by Bodnar and Wong (2003). Table 2 provides summary descriptive statistics for the exposure coefficients obtained using the eight different combinations of market and foreign exchange indices. As shown, the results are generally consistent regardless of the way the exposure coefficients are computed. All mean exposures are negative implying that on average, MNCs benefit when the dollar depreciates and vice versa. For example, the mean exposure coefficient for the nominal Broad Index using the value-weighted market control (NBVWR) is –0.5274. This indicates that, on average, a one percent increase in the dollar’s value results in a 0.53 percent decline in MNC stock prices. Approximately 30 to 40 percent of MNCs are positively exposed to the dollar’s value and the range of exposure coefficients are largest when using the broad index (both nominal and real index values). For brevity, the remaining tables in the paper report results using the Broad Index (both nominal and real values) only. It should be noted that our results are robust in all our tests to the eight alternative methods of estimating foreign exchange exposure. [Insert Table 2 about here] Table 3 describes the use of currency derivatives by sample firms. As shown in Panel A, 119 MNCs (57.2 percent of the sample) use currency derivatives. The mean level of currency derivatives by sample firms is $0.213 notional principal amount per dollar of foreign sales with the highest and lowest levels being 1.685 and 0.0005, respectively. The highest mean level is for currency forwards ($0.12 notional principal value per dollar foreign sales), followed by other currency derivatives ($0.052 notional principal value per dollar foreign sales). As shown in Panel B, the most frequently used contracts are currency forwards (used by 44.2 percent of the MNCs), followed by currency swaps (used by 14.9 percent of the MNCs) and currency options (used by 10 percent of the MNCs). It is interesting to note, that regardless of the
13
currency exposure measure used, exposure is greater for firms that do not use currency derivatives (with the exception of options users with exposure coefficients estimated using the real Broad index with the equally-weighted CRSP market index). This result is consistent with our hypothesis that currency derivatives are used to reduce the foreign exchange exposure of MNCs.13 We find statistically significant differences in mean absolute exposure coefficients for users of currency swaps and users of other foreign-exchange derivatives as compared to nonusers of derivatives. These differences are significant at the one-percent level, or better. [Insert Table 3 about here] A MNC’s operational hedging strategy is difficult to observe and measure. The ability to effectively design and implement operational hedges to reduce operating exposures is expected to be a function of a MNC’s operating flexibility. As shown by Pantzalis, Simkins, and Laux (2001), MNCs with greater breadth of operations are better equipped to effectively manage their foreign-exchange exposure. Alternatively, MNCs with networks that are concentrated in a few countries and/or regions should be less capable of effectively reducing their currency exposures. In our analysis, we use four measures to describe the MNC network of foreign operations (i.e., the breadth and depth dimensions of the network).14 Our four measures are: (1) the number of foreign countries in which the MNC has a subsidiary (NFC); (2) number of foreign regions in which the MNC operates (NFR); (3) a Herfindahl index (HERF1) measuring the concentration of an MNC network’s foreign subsidiaries across different countries; and (4) a Herfindahl index (HERF2) measuring the concentration of an MNC network’s foreign countries across different geographic regions. HERF1 and HERF2 are calculated as follows: HERF1 = Σi (NFSi)2/[ Σi (NFSi)]2 and
(4a)
HERF2 = Σj (NFCj)2/[Σ j (NFCj)]2,
(4b)
where NFSi is the number of foreign subsidiaries in country i and NFCj is the number of foreign 13
In one case the exposure coefficient is higher for users. This occurs for non-users versus options users for the RBEWR exposure coefficient.
14
Breadth describes dispersion of a firm’s operations and is measured by the number of foreign nations in which the MNC has subsidiaries. Depth measures the geographic concentration of an MNCs foreign subsidiaries. Allen and Pantzalis (1996), Doukas, Pantzalis and Kim (1999), and Pantzalis, Simkins, and Laux (2001) demonstrate the importance of breadth and depth measures in describing a MNC’s network.
14
countries in geographic region j. A higher value for either index, HERF1 or HERF2, indicates a more highly concentrated network structure and is hypothesized to be associated with increased foreign exchange exposure.15 A more geographically dispersed operational network (i.e. lower HERF1 or HERF2 index value) is a proxy for high levels of effective operational hedges and, thus, it is hypothesized to be associated with decreased foreign exchange exposure. The information related to the number of subsidiaries and their geographic location was extracted from Dunn and Bradstreet’s Who Owns Whom. This source provides a list of each firm’s domestic and foreign affiliates, and their location. These data were used in the calculation of our four measures of MNC network dispersion. The foreign subsidiaries were assigned to the following nine major geographic regions, depending on the location of each subsidiary’s host country. The nine different regions are constructed as follows: 1) The East Asia region consists of the following countries: Brunei, China, Hong-Kong, Indonesia, Japan, Korea, Macao, Malaysia, Papua-New Guinea, Singapore, Taiwan, Thailand, Vietnam, and the Philippines. 2) The NAFTA region consists of the following countries: Canada and Mexico. The U.S. is of course part of NAFTA, but U.S. subsidiaries are not counted as part of the NAFTA operations of the sample’s MNCs because they are domestic rather than foreign subsidiaries of our sample’s U.S.-based MNCs. 3) The European Union region consists of the following EU member-countries: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxemburg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom. 4) The Other Asia region includes all Asian countries not accounted for in the East Asian region. 5) The Western Europe region includes non-EU Western European countries and some Eastern European countries that have recently applied for EU membership and are expected to join the EU in the next expansion round. They are the following: Norway, Switzerland, Cyprus, Hungary, Slovenia, Poland, Czech Republic, and Malta. 6) The Eastern Europe region includes all remaining former Soviet Block and Socialist countries. 7) The Central America and Caribbean region includes all Caribbean and Central American nations.
15 For example, assume a MNC has three subsidiaries located in three different countries, HERF1 would be 2 2 2 2 0.333 [i.e. (1 + 1 + 1 )/ 3 ) = 0.333]. If all three subsidiaries are located in the same country, HERF1 is 1.0 2 2 [i.e. 3 /3 = 1.0]. HERF2 is calculated similarly. Note that the highest value for HERF1 or HERF2 is 1.0, indicating complete concentration in a single country or region.
15
8) The South America region includes all nations located in the South American continent. 9) The Africa region includes all countries of the African continent. The following section provides descriptive statistics for our operational hedges and presents cross-sectional results for examining the hypothesized relations between financial and operational hedges and currency exposure. V.
Empirical Results
A.
Descriptive Statistics Table 4, Panel A, reports descriptive statistics for four measures of foreign network
dispersion for our sample of U.S. MNCs. These measures are used as proxies of the MNC’s ability to devise operational hedges. We argue that MNCs with networks of operations that span many different countries and/or regions will be in a better position to adjust operations to effectively hedge their operating exposure. We use the variables NFC, NFR, HERF1, and HERF2, discussed in the previous section, to describe the dispersion/concentration of a MNCs foreign network. Note that higher values for NFC (the number of foreign countries in which it has operations) and NFR (the number of foreign regions in which it has operations) denote a more dispersed network. On the other hand, high values of the two alternative Herfindahl indices indicate a greater degree of concentration of a firm’s operations in a particular country (HERF1) or region (HERF2). As shown in Panel A, there are wide variations in the number of countries (NFC), number of regions (NFR), and the two Herfindahl ratios across sample firms. The average MNC in our sample operates in approximately thirteen foreign countries and four different geographic regions. The minimum and maximum number of foreign countries (foreign regions) are 1 and 47 countries (1 and 9 regions), respectively. The average Herfindahl index ranges from 0.237 (HERF1) to 0.448 (HERF2). [Insert Table 4 about here] Panel B of Table 4 compares the means of the absolute values of the estimated exposure coefficients for firms with low and high values of the foreign operations variables (e.g.,
16
NFC, NFR, HERF1, or HERF2).16 The t-tests for differences in the mean exposure for firms with high levels of dispersion in their foreign operations versus those with low dispersion suggest statistically significant differences. With the exception of the results for HERF1, which are not statistically significant, we find that firms operating in only a few countries or regions, or are highly concentrated in a particular region, have significantly higher levels of foreign-exchange exposure. This result is consistent with our earlier hypothesis that firms operating in a large number of foreign countries or regions are less exposed to foreign-exchange risk. B.
Regression Results Results of our regression analyses for Equations (2) and (3) are presented in Table 5.
Panels A and B report the determinants of absolute exposure coefficients for the dependent variables |NBEWR| (absolute value of the foreign-exchange coefficient estimated using the nominal Broad Index with EWR control) and |RBEWR| (absolute value of the foreign-exchange coefficient estimated using the real Broad Index with EWR control). In each panel, columns 1 through 4 present findings for the operational hedging measures NFC, NFR, HERF1, and HERF2, and sub-columns a and b of each column contain results for the models presented in equations (2) and (3), respectively.17 [Insert Table 5 about here] The estimated coefficients for derivatives usage (DER/FS), in our various models (1a, 2a, 3a, and 4a), range from –0.7903 to –0.9857 and are all significant at the one-percent level or better. These results suggest that the use of currency derivatives is associated with a reduction in foreign-exchange exposure. This implies that MNCs are using currency derivatives to reduce their exposure to unexpected changes in foreign-exchange rates. In addition to examining the effect of currency derivatives in the aggregate, we also examine the impact of different types of derivatives on the foreign-exchange exposure of our sample firms. These results are found in models 1b, 2b, 3b, and 4b. With the exception of currency options (OPT/FS), all other foreign-
16 A firm is assigned to the low (high) group if the value of its foreign operations variable (e.g., NFC, NFR, HERF1, or HERF2) is below (above) the sample median for that variable. 17
While not reported, the results in which the dependent variable was estimated using the CRSP valueweighted market index (NBVWR and RBVWR) are similar to those presented in Table 5.
17
exchange derivatives have negative coefficients, although the estimates for swaps (SWAP/FS) are not statistically significant. The estimated coefficients for options are generally positive but not significant in any of the models.
Both currency forwards (FWD/FS) and other foreign-
exchange derivatives (OTH/FS) have negative and statistically significant estimates. Depending on the model, the estimates for currency forwards range from –1.2053 to –1.4950 and those for other foreign-exchange derivatives range from –0.7723 to –0.9793. The estimates for these two variables are statistically significant at better than the one-percent level in all cases. These results show that the use of currency forwards and other derivatives have the effect of decreasing the firm’s foreign-exchange exposure. Neither foreign-exchange options nor swaps appear to be associated with a reduction, or an increase, in exposure. The results for the MNC network support our hypothesis that firms benefit from spreading their operations across many countries or geographic regions. The estimated coefficients for measures of the number of countries (NFC) or regions (NFR) in which a firm operates are negative and significant in all of our models. For example, the estimated coefficients for NFC (Panel A, models 1a and 2a, and Panel B, models 1a and 2a) range from –0.0373 to –0.0445 and are significant at better than the five-percent level in all estimations (the estimates are actually significant at better than the one-percent level when the dependent variable is |RBEWR|). Alternatively, our measures of concentration (e.g., HERF1 and HERF2) have significant positive estimates. Thus, firms operating in a larger number of countries or regions tend to have lower foreign-exchange exposures, while those that are concentrated in a few countries or regions tend to have higher exposures. These results are robust to various specifications of the hedging variable and currency risk exposure measures, and highlight the importance of MNC network structure in enabling the firm to develop operational hedges. We include several ownership/corporate control variables in our regressions.
These
variables are: percentage of ownership by insiders (INSDP), the square of INSDP (INSDPSQ), the percentage of ownership by blockholders (BLOCP), and the percentage of ownership by institutions (INSTP). Of these ownership/control variables, only institutional ownership (INSTP) is significant. In all models, the coefficient is positive and significant at the one percent level or
18
better. This finding is consistent with the notion that institutional investors are less concerned with foreign exchange risk, and therefore they are willing to tolerate higher exposures. Alternatively, this finding suggests that in the absence of significant institutional holdings managers have an incentive to hedge to reduce potential price pressure effects related to demand for short-selling. The estimated coefficients for leverage (LTD/TA) are negative and significant at the ten percent level or better (the estimates shown in Panel A, models 3a and 3b, and Panel B, models 1a, 1b, 3a, and 3b are significant at the five percent level or better). There are three likely explanations for this result: first, firms with higher levels of debt in their capital structures would have a higher probability of financial distress. Thus, value-maximizing managers would try to hedge to decrease the volatility of cash flows, which would reduce exposure. Second, higher monitoring intensity by debtholders, who have an incentive to reduce cash flow volatility, may cause managers to engage in hedging activities and thus reduce exposure.
Third, firms
sometimes resort to foreign-currency-denominated debt as a means of hedging. If the greater use of foreign-currency-denominated long-term debt results in greater long-term debt to assets ratios, then the negative relationship between LTD/TA and exposure could be attributed to hedging via foreign currency long-term debt financing. To summarize, we find strong evidence that the number of countries or regions, in which a firm operates, is associated with lower absolute foreign-exchange exposure.
Further,
concentration in a few countries or regions increases exposure to changes in foreign-exchange rates. We also find the use of currency derivatives, particularly forwards and other derivatives, is useful in reducing exposure. These results are robust to a variety of specifications and measures of exposure. C.
Firmwide Risk Management The regression results we present in Table 5 are consistent with the hypothesized
relationships between foreign-exchange exposure and the existence of financial and operating hedges. Next, we turn to the investigation of the coordinated use of these hedges to reduce exposure (i.e., firmwide risk management).
19
We create two variables, FIMWIDE1 and FIRMWIDE2, to explore the relationship between foreign-exchange rate exposure and coordinated (i.e., firmwide) risk management practices by MNCs. The variable FIRMWIDE1 equals zero if the number of countries in which the firm has operations (NFC) is less than the sample median and DER/FS is less than the median. Alternatively, FIRMWIDE1 equals one if either NFC or DER/FS is greater than the median. Our other measure of firmwide risk management, FIRMWIDE2, is similar except it is coded as either one or zero using the number of regions in which the firm operates (NFR). Firms operating in a small number of countries or regions and making little use of derivatives will be coded as a zero. Those MNCs operating in a larger number of countries or regions and/or using derivatives to a greater extent will be coded as a one. Thus, our firmwide measures capture coordinated riskmanagement practices by MNCs. [Insert Table 6 about here] The results for our regressions using the firmwide measures are reported in Table 6. We present results using both measures of foreign-exchange exposure (|NBEWR| and |RBEWR|) and both firmwide variables. Additionally, we include DER/FS in some of the models to see if the significant negative relationship between exposure and currency derivatives usage, reported in Table 5, persists after controlling for firmwide risk-management practices. Together, we report the results of eight separate regression models. Models 1a and 1c (2a and 2c) present the results for the NBEWR (RBEWR) exposure coefficients and firmwide measures FIRMWIDE1 and FIRMWIDE2. In all cases, the parameter estimates for the firmwide variables are negative and statistically significant at the one-percent level or better. These results strongly support the notion that MNCs using currency derivatives and operating hedges in an integrated manner can significantly reduce exposure to currency risk. Models 1b and 1d (2b and 2d) report the results for regression in which the variables HERF1 and DER/FS are also added to the model. In each of these cases, the firmwide measure is negative and significant at the ten-percent level or better. The coefficient estimates for the derivatives use and the MNC network concentration variables have the expected negative and positive sign, respectively. However, the coefficients of DER/FS and HERF1 are not significant,
20
suggesting that there is no further explanatory power gained beyond that provided by considering coordinated risk-management practices by the firm.
These results strongly support the
importance of integrated or firmwide risk-management programs by MNCs.
V.
Conclusion This study investigates the influence of financial and operating hedges on the foreign-
exchange exposure of U.S. multinational corporations. We build on previous studies of currency exposure to more fully understand how corporate risk management practices can reduce exposure to exchange-rate risk by managing risk across the firm. Our research is important for the following two reasons. First, the evidence indicates that operational hedges and financial hedges can effectively reduce foreign currency exposure. The ability to construct operational hedges, reflected in the MNC network structure, and the usage of various currency derivatives are significant determinants of currency exposure.
Furthermore, these results are robust to
alternative methods of measuring currency exposure, operating hedges, and financial hedges. Second, our results indicate that multinational corporations are taking a firmwide or “strategic” perspective in their currency risk management strategy, and are thus focusing on hedging overall economic exposure. We present strong evidence that shows the combined use of operating hedges and financial hedges is associated with decreased exchange-rate exposure. If firms were speculating using currency derivatives, or not attempting to use operational hedges effectively, we should find financial hedges and operational hedges associated with increased exposure. However, this is not the case. These results are consistent with the notion that operational and financial hedges compliment each other and support the importance of firmwide risk management in mitigating currency risk.
21
References Allen, Linda and Christos Pantzalis, 1996, Valuation of the operating flexibility of multinational corporations. Journal of International Business Studies 27: 633-653. Allayannis, George, 1996, The time-variation of the exchange rate exposure: An industry analysis, Working Paper, University of Virginia. Allayannis, George and E. Ofek, 2001, Exchange rate exposure, hedging, and the use of Foreign Currency Derivatives, Journal of International Money and Finance 20, 273-296. Allayannis, George, Jane Ihrig and James P. Weston, 2001, Exchange rate hedging: Financial vs. operational strategies, American Economic Review Papers and Proceedings 91 (2): 391-395. Banham, Russ, 2000, Ford’s better idea, Treasury and Risk Management, September 2000. Bartov, Eli and Gordon M. Bodnar, 1994, Firm valuation, earnings expectations, and the exchange rate exposure effect, The Journal of Finance 44: 1755-1785. Bartov, E., G.M. Bodnar, and A. Kaul, 1996, Exchange rate variability and the riskiness of U.S. multinational firms: Evidence from the breakdown of the Bretton Woods system, Journal of Financial Economics 42: 105-132. Belsley, D.A., E. Kuh and R.E. Welsch, 1980, Regression Diagnostics: Identifying Influential Data and Sources of Collinearity, John Wiley & Sons, Inc., New York. Bodnar, Gordon M. and W.M. Gentry, 1993, Exchange rate exposure and industry characteristics: Evidence from Canada, Japan and U.S., Journal of International Money and Finance 12: 2945. Bodnar, Gordon M. and M.H. Franco Wong, 2003, Estimating exchange rate exposures: Issues in model structure, Financial Management 32: 35-67. Bodnar, Gordon M., Gregory S. Hayt, and Richard C. Marston, 1998, 1998 Wharton survey of financial risk management by U.S.non-financial firms, Financial Management 27: 70-91. Bodnar, Gordon M., Hayt, G.S., Marston, R.C. and C.W. Smithson, 1995, Wharton survey of derivatives usage by U.S. non-financial firms, Financial Management 24: 104-114. Boehme, Rodney D., Brtley R. Danielsen and Sorin M. Sorescu, 2002, Short-sale constraints and overvaluation, University of Houston Working Paper. Booth, J.R., Smith, R.L. and Stolz, R.W., 1984, Use of interest rate futures by financial institutions, Journal of Bank Research 15: 15-20. Breeden, Douglas and S. Viswanathan, 1996, Why do firms hedge: An asymmetric information model, Fuqua School Working Paper. Buckley, Peter J. and Mark C. Casson, 1998, Models of the multinational enterprise, Journal of International Business Studies 29: 21-44. Chen, Joseph, Harrison Hong and Jeremy C. Stein, 2003, Breadth of ownership and stock returns, Journal of Financial Economics 66: 171-205. Christophe, Stephen E., 1997, Hysteresis and the value of the U.S. multinational corporation, Journal of Business 70: 435-462.
22
Choi, Jongmoo J. and Anita M. Prasad, 1995, Exchange risk sensitivity and its determinants: A firm and industry analysis of U.S. multinationals, Financial Management 24: 77-88. Chow, Edward H., Lee, Wayne Y., and Michael E. Solt, 1997a, The economic exposure of U.S. multinational firms, The Journal of Financial Research 20: 191-210. Chow, Edward H., Lee, Wayne Y., and Michael E. Solt, 1997b, The exchange-rate risk exposure of asset returns, Journal of Business 70: 105-123. Chowdhry, Bhagwan and Jonathan Howe, 1999, Corporate risk management for multinational corporations: Financial and operating hedging policies, European Financial Review 2: 229246. Cox, W. Michael, 1986, A new alternative trade-weighted dollar exchange rate index, Federal Reserve Bank of Dallas, Economic Review 67: 20-28. Davis, Henry A. and Frederick C. Militello, Jr., 1995, Foreign Exchange Risk Management: A Survey of Corporate Practices, Financial Executives Research Foundation, Morristown, New Jersey. Deragon, Jay, 2000, Old knowledge with a new name, Erisk.com, November 2000. Directory of International Affiliations, 1992, National Register Publishing Co., Skokie, Illinois. Dolde, Walter, 1995, Hedging, leverage, and primitive risk, The Journal of Financial Engineering 4: 187-216. Doukas, John, Patricia H. Hall and Larry H.P. Lang, 1999, The pricing of currency risk in Japan, Journal of Banking and Finance 23: 1-20. Doukas, John, Christos Pantzalis, and Sungsoo Kim, 1999, Intangible assets and the network structure of MNCs, Journal of International Financial Management and Accounting 10: 1-23. Dumas, Bernard and Bruno Solnik, 1995, The world price of exchange rate risk, Journal of Finance 50: 445-479. Errunza, Vihang R. and Lemma W. Senbet, 1981, The effects of international operations on the market value of the firm: Theory and evidence, Journal of Finance 36: 401-417. Errunza, Vihang R. and Lemma W. Senbet, 1984, International corporate diversification, market value, and size-adjusted evidence, Journal of Finance 39: 727-743. Figlewski , Stephen, 1981., The informational effects of restrictions on short sales: Some empirical evidence, Journal of Financial and Quantitative Analysis 16: 463-476. Flood, Eugene and Donald R. Lessard, 1986, On the measurement of operating exposure to exchange rates: A conceptual approach, Financial Management 15: 25-36. Froot, Kenneth, 1993, Currency hedging over long horizons, NBER Working Paper. Gay, Gerald D. and Jouahn Nam, 1998, The underinvestment problem and corporate derivatives use, Georgia State University Working Paper. Géczy, Christopher, Bernadette A. Minton, and Catherine Schrand, 1997, Why firms use currency derivatives, The Journal of Finance 52: 1323-1354.
23
Greene, William H., 2000, Econometric Analysis, Prentice Hall: Upper Saddle River, NJ. Jensen, Michael C., 1986, Agency costs of free cash flow, corporate finance and takeovers, American Economic Review 76: 323-329. Jensen, Michael C. and Kevin J. Murphy, 1990, CEO incentives- it’s not how much you will pay, but how, Journal of Applied Corporate Finance 3: 36-49. Jesswein, Kurt, Chuck C.Y. Kwok and William R. Folks, Jr., 1995, What new currency risk products are companies using and why?, Journal of Applied Corporate Finance 8: 103-114. Jorion, Philippe, 1990, The exchange rate exposure of U.S. multinationals, Journal of Business 63: 331-45. Jorion, Philippe, 1991, The pricing of exchange rate risk in the stock market, Journal of Financial and Quantitative Analysis 26: 363-376. Kedia, Simi, and Abon Mozumbar, 1999, Is foreign currency denominated debt a hedging instrument? Working paper Harvard University and Virginia Tech University. Kim, Wi Saeng and Esmeralda O. Lyn, 1986, Excess market value, the multinational corporation, and Tobin’s q-Ratio, Journal of International Business Studies 17: 119-125. Kogut, Bruce, 1983, Foreign direct investment as a sequential process, in C.P.Kindleberger and D. Audresch, editors, The Multinational Corporation in the 1980’s, Cambridge, Mass.: MIT Press. Kogut, Bruce, and Nalin Kulatilaka, 1995, Operating flexibility, global manufacturing, and the option value of a multinational network, Management Science 40: 123-39. Leahy, Michael P., 1998, New summary measures of the foreign exchange value of the dollar, Federal Reserve Bulletin, October, 811-818. Lehtinen, Risto, 1996, Firmwide cashflows, Firmwide Risk Management: A Risk Special Supplement, July, 30-31. Logue, Dennis, 1995, When theory fails: Globalization as a response to the (Hostile) Market for Foreign Exchange, Journal of Applied Corporate Finance 8: 39-48. McConnell, John J. and Henri Servaes, 1990, Additional evidence on equity ownership and corporate value, Journal of Financial Economics 27: 595-612. McConnell, John J. and Henri Servaes, 1995, Equity ownership and the two faces of debt, Journal of Financial Economics 39: 131-157. Mello, Antonio S., John E. Parsons, and Alexander Triantis, 1995, An integrated model of multinational flexibility and hedging policies, Journal of International Economics 33: 41-56. Mian, Shehzad L., 1996, Evidence on corporate hedging policy, Journal of Financial and Quantitative Analysis 31: 419-39. Modigliani and Miller, 1958, The cost of capital, corporation finance, and the theory of investment, American Economic Review 48: 261-297.
24
Montgomery, Douglas C. and Elizabeth A. Peck, 1992, Introduction to Linear Regression Analysis, 2nd edition, John Wiley & Sons, Inc., New York. Morck, Randall and Bernard Yeung, 1991, Why investors value multinationality, Journal of Business 64: 165-187. Nagel, Stefan, 2003, Short sales, institutitonal investors, and the book-to-market effect, London Business School Working paper. Nance, Deana R., Clifford W. Smith, Jr., and Charles W. Smithson, 1993, On the determinants of corporate hedging, Journal of Finance 48: 267-284. Pantzalis, Christos, Betty J. Simkins, and Paul Laux, 2001, Operational hedges and the foreign exchange exposure of U.S. multinational corporations, Journal of International Business Studies. Paul-Choudhury, Sumit, 1996, “Crossing the Divide”, Firmwide Risk Management, A Special Supplement to Risk July, Risk Publications, 2-4. Phillips, Aaron L., 1995, 1995 Derivatives practices and instruments survey, Financial Management 24: 115-125. Shapiro, Alan C., 1996, Multinational Financial Management, 5th edition, Prentice Hall Inc., Upper Saddle River, New Jersey. Simkins, Betty J. and Paul Laux, 1996, Derivative use and the exchange rate risk of large U.S. corporations, Conference Proceedings: The Second International Finance Conference, Georgia Tech University. Stulz, Rene M., 1988, Managerial control of voting rights: Financing policies and the market for corporate control, Journal of Financial Economics 20: 25-54. Stulz, Rene M., 1990, Managerial discretion and optimal financing policies, Journal of Financial Economics 26: 3-28. Stulz, René M., 1996, Rethinking risk management, Journal of Applied Corporate Finance 9: 824. Whidbee, David A., and Mark Wohar, 1999, Derivative activities and managerial incentives in the banking industry, Journal of Corporate Finance 5: 251-276.
25
Table 1 Description of Variables and Hypothesized Relations Between MNC Characteristics and the Absolute Value of the Foreign Exchange Exposure Coefficient This table provides a description of the variables used in the analysis. All variables are calculated for the year 1996. There are nine different geographic regions used in constructing the operational hedging variables: East Asian region, NAFTA, European Union, Other Asia region, Western Europe region, Eastern Europe region, Central America and Caribbean, South America and Africa. Hypothesis Financial Hedges
Operational Hedges
Control Variables: Size
Variable
Sign
Data Description (Source)
DER/FS
-
OPT/FS
-
FWD/FS
-
SWAP/FS
-
OTH/FS
-
NFC
-
NFR
-
HERF1
+
HERF2
+
SIZE
-
LTD/TA
-
INSDP
+
Calculated as the total notional principal amount of currency derivatives/total foreign sales. This variable includes currency derivatives such as forwards, futures, swaps, caps, collars and others. (Swaps Monitor Database) Calculated as the total notional principal amount of currency options/total foreign sales. (Swaps Monitor Database) Calculated as the total notional principal amount of currency forward contracts/total foreign sales. (Swaps Monitor Database) Calculated as the total notional principal amount of currency swaps/total foreign sales. (Swaps Monitor Database) Calculated as the total notional principal amount of currency derivatives other than options, forwards, and swaps. (i.e. includes caps, collars, floors, swaptions, etc) (Swaps Monitor Database) Number of foreign countries in which the MNC has a subsidiary. (Who Owns Whom) Number of foreign regions in which the MNC operates. (Who Owns Whom) Herfindahl Index 1. The index is calculated as follows: 2 2 HERF1 = Σ i (NFSi) /[ Σi (NFSi)] , where NFSi is the number of foreign subsidiaries in country i. Herfindahl Index 2: The index is calculated as follows: 2 2 HERF2 = Σj (NFSj) /[Σ j (NFSj)] , where NFSj is the number of foreign subsidiaries in geographic region j. Total Assets (in million dollars); (Compustat) Long-term debt/Total Assets (Compustat) Percentage of stock held by insiders. (Compact Disclosure) The squared value of the insider ownership variable
Financial Distress Insider Ownership INSDPSQ
-
Insider Ownership Squared
BLOCP
-
Blockholders
INSTP
+/-
Institutional Ownership
Percentage of stock held by blockholders. (Compact Disclosure) Percentage of stock held by institutions. (Compact Disclosure)
26
Table 2 Descriptive Statistics of Exposure Coefficients Reported are the minimum, maximum, quartiles, means, standard deviations and ranges for the following eight alternative foreign exchange exposure coefficients: NBEWR is the exposure coefficient obtained using the nominal broad exchange rate index and an equally weighted market portfolio. NBVWR is the exposure coefficient obtained using the nominal broad exchange rate index and a value weighted market portfolio. NMEWR is the exposure coefficient obtained using the nominal major currencies’ exchange rate index and an equally weighted market portfolio. NMVWR is the exposure coefficient obtained using the nominal major currencies’ exchange rate index and a value weighted market portfolio. NOEWR is the exposure coefficient obtained using the nominal other currencies’ exchange rate index and an equally weighted market portfolio. NOVWR is the exposure coefficient obtained using the nominal other currencies’ exchange rate index and a value weighted market portfolio. RBEWR is the exposure coefficient obtained using the real broad exchange rate index and an equally weighted market portfolio. RBVWR is the exposure coefficient obtained using the real broad exchange rate index and a value weighted market portfolio. All exposure coefficients (γi) were obtained from the following time series regression: Rit = αi + βiRmt + γiRct + εit,where Ri,t is the return of stock i in month t, RM,t is the return on the market portfolio in month t, and RC,t is the percent change in the exchange rate index in month t. Regressions were run over a 60-month period covering the years 19941998. [N=208]
NBEWR
NBVWR
NMEWR
NMVWR
NOEWR
NOVWR
RBEWR
RBVWR
min
-5.8799
-6.0626
-4.4696
-4.5892
-3.0903
-3.1192
-6.0193
-6.3199
Q1
-0.8494
-0.9176
-0.7709
-0.7740
-0.4920
-0.5288
-0.8496
-1.0045
median
-0.4335
-0.3907
-0.2884
-0.2781
-0.0967
-0.0957
-0.3452
-0.3989
Q3
0.0906
0.1357
0.1885
0.2213
0.1135
0.1299
0.1611
0.1350
max
2.1961
2.1829
2.2099
1.7572
1.3889
1.2707
2.1578
2.2469
mean
-0.0508
-0.5274
-0.3225
-0.3195
-0.2418
-0.2484
-0.5002
-0.5901
std. dev.
1.1117
1.1620
0.9392
0.9556
0.6575
0.6795
1.1376
1.2028
positive
31.25 %
31.73 %
35.10 %
37.50 %
38.46 %
39.90 %
30.29 %
28.85 %
range
8.076
8.2455
6.6795
6.3464
4.4792
4.3899
8.1771
8.5668
27
Table 3 Currency Derivatives Usage by Sample Firms Panel A: Descriptive statistics of currency derivatives’ notional values as a percent of foreign sales for the subsample of the MNC firms that use currency derivatives. Definitions are as follows: DER/FS = Total currency derivatives notional value / foreign sales, OPT/FS = currency options notional value / foreign sales, FWD/FS = currency forwards notional value / foreign sales, OTH/FS = other currency derivatives notional value / foreign sales. [N=119]
DER/FS
OPT/FS
FWD/FS
SWAP/FS
OTH/FS
min
0.0005
0
0
0
0
Q1
0.0475
0
0.0082
0
0
median
0.1249
0
0.0706
0
0
Q3
0.3129
0.0497
0.1406
0.0049
0
max
1.6851
0.2193
0.8890
1.1088
1.5094
mean
0.2130
0.0142
0.1197
0.0350
0.0517
std. dev.
0.2786
0.0447
0.1641
0.1180
0.1834
Panel B: Mean values of absolute exposures for non-users and users of derivatives. |NBEWR| is the currency exposure coefficient obtained using the nominal Broad Index and the equally-weighted CRSP market portfolio return. |NBVWR| is the currency exposure coefficient obtained using the nominal Broad Index and the value-weighted CRSP market portfolio return. |RBEWR| is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the equally-weighted CRSP market portfolio return. |RBVWR| is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the value-weighted CRSP market portfolio return. Note, *, **, and *** denote that the mean value is significantly different than that of the non-users’ group at the 10%, 5% and 1% significance level, respectively. Non-Users
Users
89
All Users 119
Options Users 21
Forwards Users 92
Swaps Users 31
Other derivatives users 30
42.79 %
57.21 %
10.01%
44.23 %
14.90 %
14.42 %
|NBEWR|
0.9073
0.7388
0.8911
0.7918
0.5158 ***
0.5107 ***
|NBVWR|
0.9690
0.7581
0.8381
0.8077
0.4614 ***
0.5514 ***
|RBEWR|
0.8975
0.7740
0.9122
0.8342
0.5406 ***
0.5388 ***
|RBVWR|
0.9866
0.8201
0.8781
0.8772
0.5085 ***
0.5921 ***
N % of all firms
28
Table 4 Foreign Operations Network of Sample Firms The foreign operations network variables are used as proxies for the firms ability to construct operational hedges for foreign exchange risk management. Descriptive statistics (Panel A) and comparisons of mean exposures of different groups of MNCs classified based on whether their foreign operations network variables are above or below the sample median. The following variables are examined: NFC = number of foreign countries the MNC is operating in. NFR = number of foreign regions the MNC operates in. HERF1 is a concentration (Herfindahl) ratio of the number of 2 2 foreign subsidiaries across foreign countries, defined as HERF1 = Σi (NFSi) /[Σi (NFSi)] , where NFSi is the number of foreign subsidiaries in country i. HERF2 is a concentration (Herfindahl) ratio of the number of foreign countries 2 2 across foreign regions, defined as HERF2 = Σj (NFCj) /[Σj (NFCj)] , where NFCj is the number of foreign countries in geographic region j. There are nine different geographic regions: Asian Crisis region, NAFTA, European Union, Other Asia region, Western Europe region, Eastern Europe region, Central America and Caribbean, South America and Africa. Panel A: Descriptive statistics for variables describing the foreign operations network of MNCs. [N=208]
NFC
NFR
HERF1
HERF2
min
1
1
0.0424
0.1728
Q1
5
2
0.0944
0.2717
median
11
4
0.1481
0.3574
Q3
20
6
0.2658
0.52
max
47
9
1
1
mean
13.35
4.22
0.2369
0.4477
Std. dev
10.43
2.13
0.2359
0.2521
Panel B: Comparison of absolute exposures’ mean values for firms with “low” and “high” values of foreign operations variables. A firm is assigned to the “low” (“high”) group if the value of the particular variable for the firm is below (above) the sample median value. Note, *, **, and *** denote that the mean value for the “High” group is significantly different than that of the “Low” group at the 10%, 5% and 1% significance level, respectively. |NBEWR|, is the currency exposure coefficient obtained using the nominal Broad Index and the equally-weighted CRSP market portfolio return. |NBVWR |, is the currency exposure coefficient obtained using the nominal Broad Index and the value-weighted CRSP market portfolio return. |RBEWR|, is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the equally-weighted CRSP market portfolio return. |RBVWR|, is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the value-weighted CRSP market portfolio return. NFC
NFR
HERF1
HERF2
“Low”
“High”
“Low”
“High”
“Low”
“High”
“Low”
“High”
|NBEWR|
0.9520
0.6585 **
0.9561
0.6629 **
0.7235
0.8983
0.6336
0.9882 ***
|NBVWR|
1.0096
0.6742 ***
1.0246
0.6686 ***
0.7498
0.9469
0.6415
1.0552 ***
|RBEWR|
0.9801
0.6613 **
0.9738
0.6770 **
0.7397
0.9140
0.6554
0.9983 ***
|RBVWR|
1.0693
0.6992 ***
1.0783
0.7008 ***
0.7888
0.9939
0.6756
1.1071 ***
29
Table 5 Relationship Between Absolute Exposure and Financial and Operational Hedges This table reports estimates of the determinants of the absolute exchange rate exposure. Weighted least squares regressions with robust standard errors (White (1980)) are used to control for heteroskedasticity. The weighted least squares procedure weights by the exposure coefficient’s standard error. Significance levels are indicated as follows: *** 1%, ** 5%, * 10%. The independent variables are: SIZE = Total assets. LTD/TA = Long term debt to total assets ratio. INSDP = %-shareholdings by insiders. INSDPSQ = INSDPsquared. BLOCP = %-shareholdings by blockholders. INSTP = %-shareholdings by institutions. MNC network = NFC or NFR or HERF1 or HERF3. DER/FS = Derivatives’ notional value / Foreign sales. OPT/FS = Options notional value / Foreign sales. FWD/FS = Forwards notional value / Foreign sales. SWAP/FS = Swaps notional value / Foreign Sales. OTH/FS = Other derivatives notional value / Foreign sales. Panel A: The dependent variable, |NBEWR|, is the currency exposure coefficient obtained using the nominal Broad Index and the equally-weighted CRSP market portfolio return. Variable
MNC network (operational hedging ability) is measured by : NFC 1a
Intercept SIZE LTD/TA INSDP INSDPSQ BLOCP INSTP MNC network DER/FS
0.4963 (1.12) -6 2.6×10 (1.07) -1.6458 * (-1.80) 0.0767 (1.49) -0.0014 (-1.39) -0.0112 (-1.32) 0.0251 *** (2.85) -0.0373 ** (-2.37) -0.8186 *** (-3.14)
208 0.2451 3.01
FWD/FS SWAP/FS OTH/FS
Adj.-R
2
F-value
1b 0.5424 (1.21) -6 2.1×10 (0.88) -1.6777 * (-1.81) 0.0759 (1.50) -0.0014 (-1.34) -0.0117 (-1.36) 0.0250 *** (2.83) -0.0389 ** (-2.38)
2.0734 (0.83) -1.2053 *** (-2.72) -0.4394 (-0.66) -0.8145 *** (-3.37) 208
OPT/FS
N
NFR 2a 0.6459 (1.50) -6 1.3×10 (0.57) -1.4955 * (-1.75) 0.0758 (1.51) -0.0014 (-1.40) -0.0107 (-1.30) 0.0255 *** (2.87) -0.1636 ** (-2.51) -0.8120 *** (-3.41)
HERF1 2b
0.6913 (1.58) -6 0.5×10 (0.25) -1.5120 * (-1.76) 0.0755 (1.49) -0.0014 (-1.36) -0.0112 (-1.34) 0.0254 *** (2.85) -0.1683 ** (-2.54)
208
1.7527 (0.73) -1.3112 *** (-2.89) -0.3843 (-0.63) -0.7909 *** (-3.58) 208
0.2479
0.2495
3.19
2.98
3a -0.1112 (-0.21) -6 -0.7×10 (-0.31) -1.9382 ** (-2.18) 0.0748 (1.63) -0.0015 (-1.53) -0.0078 (-1.12) 0.0239 *** (2.85) 1.0320 * (1.75) -0.9758 *** (-3.60)
HERF2 3b
-0.0892 (-0.17) -6 -1.7×10 (-0.84) -1.9604 ** (-2.17) 0.0753 (1.62) -0.0015 (-1.51) -0.0083 (-1.17) 0.0240 *** (2.83) 1.0342 * (1.74)
4a -0.5314 (-0.80) -6 0.2×10 (0.08) -1.5156 * (-1.88) 0.0769 (1.58) -0.0015 (-1.49) -0.0098 (-1.25) 0.0253 *** (2.86) 1.1801 ** (2.11) -0.9842 *** (-3.61)
4b -0.5106 (-0.76) -6 -0.8×10 (-0.37) -1.5326 * (-1.87) 0.0772 (1.57) -0.0015 (-1.47) -0.0102 (-1.29) 0.0253 *** (2.84) 1.1870 ** (2.11)
208
-0.0831 (-0.04) -1.3757 *** (-3.02) -0.0682 (-0.10) -0.9696 *** (-4.13) 208
208
0.4658 (0.21) -1.4386 *** (-3.05) -0.3236 (-0.52) -0.8798 *** (-4.09) 208
0.2521
0.2396
0.2416
0.2496
0.2516
3.19
2.64
2.83
2.76
3.02
30
Table 5 cont’d. Panel B: The dependent variable, |RBEWR|, is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the equally-weighted CRSP market portfolio return. Variable
MNC network (operational hedging ability) is measured by : NFC
Intercept SIZE LTD/TA INSDP INSDPSQ BLOCP INSTP MNC network DER/FS
1a
1b
2a
2b
3a
0.5318 (1.09) -6 3.1×10 (1.26) -1.6410 ** (-1.99) 0.0781 (1.52) -0.0012 (-1.23) -0.0135 (-1.59) 0.0268 *** (3.03) -0.0445 *** (-2.71)
0.6461 (1.41) -6 2.1×10 (0.91) -1.4335 * (-1.86) 0.0783 (1.58) -0.0013 (-1.29) -0.0123 (-1.53) 0.0274 *** (3.08) -0.1859 ** (-2.87) -0.8403 *** (-3.24)
0.7000 (1.50) -6 1.3×10 (0.59) -1.4607 * (-1.87) 0.0777 (1.56) -0.0012 (-1.25) -0.0129 (-1.57) 0.0273 *** (3.06) -0.1920 *** (-2.91)
-0.2079 (-0.36) -6 -0.2×10 (-0.08) -1.9273 ** (-2.24) 0.0777 * (1.71) -0.0014 (-1.48) -0.0091 (-1.32) 0.0256 *** (3.01) 1.1440 * (1.86) -0.9727 *** (-3.54)
208
2.6743 (1.01) -1.2290 *** (-2.70) -0.4376 (-0.61) -0.7982 *** (-2.89) 208
0.2628 3.34
FWD/FS SWAP/FS OTH/FS
Adj.-R
2
F-value
HERF1
0.4771 (1.00) -6 3.7×10 (1.46) -1.6042 ** (-1.97) 0.0792 (1.56) -0.0013 (-1.28) -0.0129 (-1.55) 0.0270 *** (3.05) -0.0425 *** (-2.70) -0.7903 *** (-2.98)
OPT/FS
N
NFR
208
2.2925 (1.01) -1.3506 *** (-2.97) -0.3740 (-0.58) -0.7723 *** (-3.00) 208
0.2669
0.2679
2.96
3.28
HERF2 3b
-0.1830 (-0.32) -6 -1.3×10 (-0.59) -1.9528 ** (-2.24) 0.0781 * (1.70) -0.0014 (-1.46) -0.0097 (-1.37) 0.0256 *** (2.99) 1.1472 * (1.86)
4a -0.6392 (-0.89) -6 0.7×10 (0.29) -1.4588 * (-1.93) 0.0810 * (1.69) -0.0014 (-1.44) -0.0112 (-1.45) 0.0269 *** (3.00) 1.2454 ** (2.13) -0.9857 *** (-3.57)
4b -0.6163 (-0.85) -6 -0.3×10 (-0.14) -1.4780 * (-1.93) 0.0814 * (1.68) -0.0014 (-1.41) -0.0117 (-1.50) 0.0269 *** (2.97) 1.2541 ** (2.12)
208
0.1781 (0.09) -1.4267 *** (-3.10) -0.0160 (-0.02) -0.9793 *** (-3.92) 208
208
0.7311 (0.36) -1.4950 *** (-3.22) -0.2928 (-0.44) -0.8877 *** (-3.55) 208
0.2717
0.2534
0.2561
0.2600
0.2628
2.94
2.90
2.67
3.05
2.80
31
Table 6 Relationship Between Absolute Exposure and Firmwide Risk Management This table reports the relationship between absolute exchange rate exposure and the combined usage of financial and operating hedges, our test for firmwide risk management. Weighted least squares regressions with robust standard errors (White (1980)) are used to control for heteroskedasticity. The weighted least squares procedure weights by the exposure coefficient’s standard error. Significance levels are indicated as follows: *** 1%, ** 5%, * 10%. The independent variables are: SIZE = Total assets. LTD/TA = Long term debt to total assets ratio. INSDP = %-shareholdings by insiders. INSDPSQ = INSDP-squared. BLOCP = %-shareholdings by blockholders. INSTP = %-shareholdings by institutions. We use two indicator measures of combined hedging, FIRMWIDE1 and FIRMWIDE2. FIRMWIDE1 takes the value of 2 if both NFC and DER/FS are larger than the corresponding sample medians, the value of 1 if one of the two variables’ value is greater than the median, and the value of 0 otherwise. FIRMWIDE2 is similar to FIRMWIDE1 but NFR is used instead of NFC. Dependent variable = |NBEWR|
Dependent variable = |RBEWR|
Variable FIRMWIDE1
1a
FIRMWIDE2
1b 0.1411 (0.24)
1c 0.3141 (0.72)
FIRMWIDE1
1d 0.1414 (0.25)
2a 0.3512 (0.76)
FIRMWIDE2
2b 0.1356 (0.23)
2c
Intercept
0.3747 (0.84)
SIZE
-2.01×10 (-0.11)
LTD/TA
-1.6219 * (-1.80)
-1.8480 ** (-2.10)
-1.4437 * (-1.68)
-1.6716 * (-1.94)
-1.5278 ** (-1.99)
-1.8038 ** (-2.17)
-1.3561 * (-1.79)
INSDP
0.0776 (1.46)
0.0692 (1.53)
0.0791 (1.52)
0.0718 (1.59)
0.0787 (1.52)
0.0701 (1.58)
0.0807 (1.60)
INSDPSQ
-0.0013 (-1.32)
-0.0013 (-1.35)
-0.0013 (-1.35)
-0.0013 (-1.37)
-0.0012 (-1.20)
-0.0011 (-1.24)
-0.0012 (-1.20)
BLOCP
-0.0103 (-1.24)
-0.0092 (-1.28)
-0.0094 (-1.19)
-0.0088 (-1.24)
-0.0120 (-1.47)
-0.0110 (-1.54)
-0.0120 (-1.47)
INSTIT
0.0243 *** (2.81)
0.0248 *** (2.96)
0.0253 *** (2.92)
0.0255 *** (3.03)
0.0263 *** (3.09)
0.0267 *** (3.23)
0.0275 *** (3.23)
FIRMWIDE
-0.5365 *** (-3.17)
-0.3651 * (-1.73)
-0.5666 *** (-3.27)
-0.4219 * (-1.89)
-0.6346 *** (-3.74)
-0.4963 ** (-2.38)
-0.6642 *** (-3.88)
208
208
208
208
208
208
208
0.2437
0.2459
0.2550
0.2631
0.2725
0.2832
0.2858
0.2919
2.85 ***
2.67 ***
3.09 ***
3.02 ***
3.69 ***
3.22 ***
3.72 ***
3.29 ***
-7
2.59×10 (0.14)
-7
-0.01×10 (-0.01)
-7
3.42×10 (0.19)
-7
9.80×10 (0.53)
-7
11.00×10 (0.63)
0.6739 (1.01)
0.5453 (0.78)
0.6562 (0.96)
DER/FS
-0.3974 (-1.32)
-0.3276 (-1.13)
-0.1855 (-0.58)
Adjusted-R F-value
2
2d
0.1246 (0.21) 11.70×10-7 (0.68) -1.5758 * (-1.90) 0.0739 * (1.68) -0.0011 (-1.27) -0.0104 (-1.48) 0.0276 *** (3.35) -0.5545 ** (-2.53) 0.5029 (0.71) -0.1196 (-0.38) 208
HERF1
N
0.2773 (0.62) -7
11.60×10 (0.64)
-7
Appendix 1 Currency Indexes and Trade Weights The table presents the 1997 trade weights for the U.S. dollar indexes used in the study. Index weights are updated annually and previous years index weights are available at the Federal Reserve Board’s web site (http://www.federalreserve.gov). Weights are listed in percent. Components may not sum to totals because of rounding. Country or Region Canada Euro area (see below) Japan Mexico China United Kingdom Taiwan Korea Singapore Hong Kong Malaysia Brazil Switzerland Thailand Australia Indonesia Philippines Russia India Sweden Saudia Arabia Israel Argentina Venezuela Chile Columbia
Broad Index 17.3 16.4 14.6 8.6 6.6 4.6 3.9 3.7 3.1 2.8 2.4 1.9 1.8
Major Index 30.3 28.7 25.6
OITP Index
19.9 15.3 8.0 9.1 8.6 7.2 6.6 5.5 4.4 3.2
1.7 1.5 1.3 1.2 0.9 0.9 0.9 0.9 0.9 0.6 0.6 0.5 0.5
3.9 2.6 3.0 2.7 2.2 2.2 1.6 2.1 2.1 1.5 1.4 1.3 1.1
Total
100.0
100.0
Euro-area countries Germany France Italy Netherlands Belgium/Luxenbourg Spain Ireland Austria Finland Portugal Total Euro-area
5.6 2.9 2.5 1.5 1.4 0.8 0.7 0.4 0.3 0.1 16.4
9.9 5.0 4.5 2.7 2.5 1.4 1.3 0.7 0.6 0.2 28.7
33
100.0
Department of Finance, College of Business Administration, Oklahoma State University, Stillwater, OK 74078-4011, USA b Department of Finance, College of Business Administration, University of South Florida, Tampa, Florida 33620-5500, USA Version of April 28, 2003 Abstract This paper investigates the impact of firmwide risk management practices on the foreign exchange exposure of 208 U.S. multinational corporations (MNC) over the period 1994 to 1998. Firmwide risk management is referred to here as the coordinated use of both financial hedges, such as currency derivatives, and operational hedges, described by the structure of a firm’s MNC foreign subsidiary network. We find that the use of currency derivatives, particularly forward contracts, is associated with reduced levels of foreign-exchange exposure. Furthermore, MNCs with dispersed operating networks have lower levels of currency exposure. These findings are robust to alternative ways of measuring foreign exposure. Finally, our results strongly support the view that MNCs hedging in a coordinated manner can significantly reduce exposure to currency risk. These results strongly suggest that operational and financial hedges are complementary risk management strategies.
JEL Classification: F30, F31, G15 Key words: Multinational finance, Risk management, Firmwide risk management
* Please direct all correspondence to: Betty J. Simkins, Department of Finance, College of Business Administration, Oklahoma State University, Stillwater, OK 74078-4011, (405) 744-8625 (office), (405) 744-5180 (fax), email: [email protected].
We thank Cesare Robotti for comments on an earlier draft of this paper. A prior version of this paper was presented at the 2000 Financial Management Association meetings.
Firmwide Risk Management of Foreign Exchange Exposure by U.S. Multinational Corporations Abstract This paper investigates the impact of firmwide risk management practices on the foreign exchange exposure of 208 U.S. multinational corporations (MNC) over the period 1994 to 1998. Firmwide risk management is referred to here as the coordinated use of both financial hedges, such as currency derivatives, and operational hedges, described by the structure of a firm’s MNC foreign subsidiary network. We find that the use of currency derivatives, particularly forward contracts, is associated with reduced levels of foreign-exchange exposure. Furthermore, MNCs with dispersed operating networks have lower levels of currency exposure. These findings are robust to alternative ways of measuring foreign exposure. Finally, our results strongly support the view that MNCs hedging in a coordinated manner can significantly reduce exposure to currency risk. These results strongly suggest that operational and financial hedges are complementary risk management strategies.
I.
Introduction The practice of corporate risk management has changed dramatically over the past two
decades. Originally, risk management was implemented on an uncoordinated basis across different units of the firm. The primary focus of these ad hoc risk management programs was to minimize costs of particular units. Today, however, risk management of currency exposure has, in many cases, evolved into a firmwide exercise that addresses both short-term and long-term exposures and encompasses financial as well as operational hedges. The ultimate goal of firmwide risk management is to reduce risk while placing the firm in a position to benefit from opportunities that arise from exchange rate changes.1 For example, Davis and Militello (1995) describe how Union Carbide employs a firmwide perspective in risk management. The company uses a one-year horizon for financial hedges (e.g., foreign-exchange derivatives), whereas for longer horizons, operational adjustments are made in sourcing, utilization of different plant locations, and pricing. We define firmwide risk management for multinational corporations (MNCs) as the combined use of both financial and operational hedges as part of an integrated risk management strategy aiming at reducing exposure to foreign-exchange risk.2 Changes in exchange rates can
1
See Paul-Choudhury, 1996.
2
Firmwide risk management is one of the most widely used out of a group of synonyms that describe a broad and comprehensive view of managing risk across the firm. Other terms used to describe this type of coordinated risk management are enterprise risk management, global risk management, and strategic risk
1
influence MNCs’ current and future expected cash flows and ultimately, stock prices. The direction and the magnitude of changes in the exchange rate on firm value are a function of the firm’s corporate hedging policy and the structure of its foreign currency cash flows. The latter depends on the firm’s competitive position in the industries in which it operates. The former indicates whether the MNC utilizes operational hedges and financial hedges to manage currency exposure. Beginning with the seminal study by Jorion (1990), initial research in this area focused on whether corporations are exposed to foreign exchange risk (see Bodnar and Gentry, 1993, Bartov and Bodnar, 1994, 1995, and Chow, Lee and Solt, 1997a,b). Allayannis and Ofek (2001) and Simkins and Laux (1996) investigate the effect of financial hedging on foreign-exchange exposure. More recently, Pantzalis, Simkins, and Laux (2000) examine the ability of operational hedges to reduce exposure.
However, few studies thus far have examined the combined
influence of financial hedges and operational hedges on foreign exchange exposure.3 Our research fills this gap in the literature by examining the relationship of integrated financial and operational hedging programs (firmwide risk management) and foreign exchange exposure for a sample of U.S.-based MNCs. It is common practice among firms to use a combination of production and marketing strategies across the firm’s different operating units (operational hedges) to manage long term exposure, whereas foreign exchange derivatives (financial hedges) are more often used for managing short term exposure. Long-term operating policy adjustments are costly and difficult to reverse, hence they are most effective when the firm possesses a network of multiple operating units that span many business and geographic areas. We describe the firm’s ability to implement operational hedges by examining the network structure of its foreign operations (i.e., the MNC
management. Firmwide risk management addresses exposures on a global basis, in that it considers all parts of the firm and tries to cope with both short-term and long-term exposures. For example, The Tower Group uses the term enterprise risk management and defines it as “the process of managing the risk faced by an institution on a global, institution-wide basis.” (see Deragon, 2000). As reported by Banham (2000), Freeman Wood, Ford Motor Company’s director of global risk management, states: “We want to understand our total aggregate exposure to financial market risks and nonfinancial market risks. The next phase is measuring these risks in such a way as to permit discussion and comparison. That is followed by a determination of which risks to retain, hedge or transfer. Finally, we’ll determine the optimum risk transfers.”
2
network).4 We analyze financial hedges by utilizing information on currency derivatives’ usage. We consider usage of different foreign currency derivatives contracts such as futures, forwards, swaps, options, forward rate agreements, and other derivative contracts that have a foreign currency as the underlying commodity. We argue that the coordinated use of operational and financial hedges should more effectively reduce exposure to foreign exchange risk, because it addresses the firm’s overall exposure both in the short- and in the long-term. The two important contributions of our research are as follows.
First, the evidence
suggests that the coordinated use of operational and financial hedges can effectively reduce foreign currency exposure. To our knowledge, this is the first study to empirically document this result.
Furthermore, these results are robust to alternative methods of measuring currency
exposure, operating hedges, and financial hedges. Second, our results indicate that multinational corporations are taking a firmwide or “strategic” perspective in their currency risk management strategy, and are thus focusing on hedging overall economic exposure. This paper proceeds as follows: Section II examines prior research on foreign-exchange exposure and the use of operational and financial hedging to manage exposure. Section III discusses the methodology we use to investigate the determinants of foreign-exchange exposure by U.S. multinational firms. Our data sources and sample used are described in Section IV, while Section V reports our empirical results. We conclude the paper in Section VI.
II.
Multinational Corporations, Foreign Exchange Exposure, and Firmwide Risk Management A number of studies have examined the uniqueness of U.S. multinational corporations
(MNCs). For example, Errunza and Senbet (1981, 1984), Kim and Lyn (1986), and Morck and Yeung (1991) investigate the value of international operations. However, as Christophe (1997) points out, these studies generally focus on MNC activities during the 1970s, a period
3
The only two previous studies that examined operational as well as financial hedges are Allayannis et al (2001) and Allayannis and Weston (2002). Their findings are discussed in the next section. 4 An MNC’s network has to do with the number of different foreign nations in which operations are conducted (“breadth” of MNC network), as well as the concentration of operations in various locations (“depth” of MNC network).
3
characterized by relatively stable exchange rates. In fact, the exchange rate volatility that has existed since the 1970s makes it necessary to investigate the means by which MNCs manage their risk, and the effectiveness of their risk-management techniques. Financial and operational hedges are two fundamental ways MNCs can manage the risk introduced by increased exchange rate volatility. Flood and Lessard (1986) identify two types of exposure to foreign exchange risk: 1) transaction exposure and 2) operating exposure. Transaction exposure is the effect of unanticipated changes in real exchange rates on nominal cash flows (i.e., cash flows fixed in nominal terms) and primarily a short-term exposure that can be hedged using financial derivatives. In contrast, operating exposure is the effect of unanticipated changes in exchange rates on the cash flows associated with a firm’s real assets and liabilities and is, therefore, primarily a long-term exposure that amounts to the impact of unexpected changes in the exchange rate on the firm’s competitive position. Logue (1995) and Chowdhry and Howe (1999) argue that operating exposure cannot be effectively managed using financial hedges. Instead, they suggest that long-term strategy adjustments (i.e., operational hedges) are the most effective way of managing long-run operating exposure. A firm facing future, contractually fixed, foreign-currency cash flows, in which the only source of uncertainty is the exchange rate (i.e., transaction exposure), can easily hedge with swaps or forward contracts.5 However, if the future cash flows are uncertain and not perfectly correlated with the exchange rate (i.e., operating exposure), financial hedging is likely to be ineffective. A firm’s operating exposure to currency risk depends on the effect of unexpected changes in the exchange rate on the firm’s output prices (e.g., product prices) and input costs (e.g., raw materials, labor costs, etc.). Since the correlation of prices with exchange rates is determined by the degree of segmentation of their respective markets, operating exposure depends on whether input costs and output prices are determined locally or globally. Firms
5 In fact, annual surveys conducted by the Wharton School and Chase Manhattan Bank find that many firms use currency derivatives to manage foreign-exchange risk (see Bodnar, Hayt, and Marston, 1998). Surveys of corporate treasurers also reflect a trend toward management, not only of transaction exposure, but also translation exposure and economic exposure broadly construed (Jesswein, Kwok, and Folks, 1995).
4
facing substantial operating exposures, as is often the case for U.S. MNCs, can manage this exposure by devising operating strategies that consist of combinations of different marketing initiatives, such as market selection or pricing strategy, and production initiatives, such as raw materials sourcing and production location (see Shapiro, 1996). Thus, operational hedges entail long-term operating policy adjustments that are implemented within a firm’s network of operating units. Theoretically, the effectiveness of these policies in managing operating exposures is enhanced when they can be implemented across different lines of business and locations.6 Obviously, not every firm will be able to manage currency risk exposure with the same degree of effectiveness. We argue that firms with operations spread over many currency and business areas are more insulated from foreign exchange exposure, since they have at their disposal a larger set of alternatives to devise effective operational hedges. In effect, because of their geographic diversification, MNCs own a portfolio of real options not available to purely domestic firms. This operating flexibility allows MNCs to effectively react to exchange rate changes by deciding where to shift production, source for inputs, or declare profits among the different locations in which they operate.7 The degree of operating flexibility of a MNC is clearly a function of the structure of its foreign operations’ network. We postulate that MNCs that have operations in a large number of geographic regions will be in a better position to effectively construct operational hedges than similar-sized MNCs whose operations are concentrated in few geographic regions. Most recent empirical studies in the risk management area have examined financial hedging by nonfinancial corporations, including MNCs.
For instance, Nance, Smith, and
Smithson (1993), Dolde (1995), and Mian (1996) examine the hedging policies of corporations. Further, Géczy, Minton, and Schrand (1997) investigate the use of currency derivatives by firms. They find that corporations with extensive exposure to foreign exchange risk are more likely to 6
For example, Cummins Engine Company states they have substantially reduced exposure by building manufacturing plants in three major economic regions of the world – the U.S., Europe, and Japan. National Semiconductor says they take into account economic exposure when selecting manufacturing plant locations (see Davis and Militello, 1995). 7
For a discussion of the MNCs’ operating flexibility see Kogut (1983), Kogut and Kulatilaka (1995), and Buckley and Casson (1998) among others. For empirical evidence regarding the value of operating flexibility see Allen and Pantzalis (1996).
5
use currency derivatives. While these studies focus on the determinants of a firm’s hedging policies, they do not investigate the effectiveness of the policies. To date, only a few studies have investigated the effect of financial hedging on foreign-exchange exposure (see, for example, Allayannis and Ofek, 2001, and Simkins and Laux, 1997). Further, recently there has been more interest in the use of operational hedges by MNCs seeking to reduce exposure (see Pantzalis, Simkins, and Laux, 2001). Most importantly, though, is the fact that there has been very limited empirical evidence on the combined use of financial and operational hedges in foreign exchange risk management. The only study that has provided some preliminary evidence is by Allayannis et al (2001). In their study, Allayannis et al utilize measures of the MNC network (similar to the ones found in Pantzalis et al (2001)) as proxies for the firms’ ability to implement operational hedges. They find that operational hedges are not effective for risk management. Furthermore, they find that the likelihood of financial derivatives’ use increases with the geographic dispersion of the U.S. MNCs’ operations and that operational strategies benefit shareholders when used in combination with financial strategies. However, they restrict their tests to firms with positive currency exposure coefficients and do not control for size and several other factors that may affect exposure. Our research fills this gap in research by examining firmwide risk management of MNCs. Obviously, managing risk across a global firm is a complex task, not only because exposures may be difficult to quantify but also because hedging a group of exposures typically requires a more sophisticated approach than hedging individual exposures. With the development of sophisticated risk management techniques and digital data transfer technology, companies are now able to effectively evaluate and manage firmwide cashflows on a global basis, by incorporating financial and operational hedges into an integrated analysis (Lehtinen, 1996). Proponents of a firmwide approach to risk management argue that potential benefits from it can be profound.
By viewing decisions related to financial transactions and operations as
interrelated, risk can be reduced, controlled or transferred to minimize variations that affect performance.
6
III.
Methodology
A.
Measuring Foreign Exchange Exposure We use a two-factor market model to estimate the effect of exchange rate changes on
the stock returns for our sample firms. This methodology is standard in other research examining currency exposure (e.g., Jorion, 1990, Bodnar and Gentry, 1993, Bartov and Bodnar, 1994, Choi and Prasad, 1995, Simkins and Laux, 1997, and Bodnar and Wong, 2003). For each firm in our sample, we estimate the following time-series regression: Rit = αi + βiRmt + γiRct + εit,
(1)
where Rit is the rate of return on the ith company’s common stock in month t, Rmt is the return on the market index, Rct is the return on the exchange rate index, and εit is the idiosyncratic error term.
The exchange rate is measured as the price of the U.S. dollar in foreign currency;
therefore, a positive value for Rct indicates a strengthening dollar. For each firm, the estimated coefficient γi is a measure of the sensitivity of a firm’s stock returns to changes in foreignexchange rates, and can be either positive or negative. We estimate the stock-price sensitivity to changes in exchange rates for our sample firms for the period 1994-98. This sample period is chosen for two reasons. First, we know which firms actively use currency derivatives over this period based on reporting disclosures and limit our study to a period where hedging practices are most likely to be consistent. Second, prior studies by Bartov and Bodnar (1995), Allayannis (1996), and Chow, Lee, and Solt (1997a) find that a firm’s exposure changes over longer time frames. Among other factors, this is likely due to changes in a firm’s MNC network. To avoid time-varying exposure coefficients, we limit our study to a five-year period. B.
Determinants of Foreign Exchange Exposure
We hypothesize that financial and operational hedges can be effective in reducing foreign exchange exposure. To test this hypothesis, we analyze the following cross-sectional regression: ⎮γ i⎮ = δ0 + δ1SIZEi + δ2LTD/TAi + δ3INSDPi + δ4INSDPSQi + δ5BLOCPi + δ6INSTPi + δ7MNCi + δ8DER/FSi + εi,
7
(2)
where ⎮γi⎮ is the absolute value of the ith firm’s foreign exchange exposure coefficient estimated in Equation (1).8 Since, as discussed earlier, firms can be either negatively or positively exposed to currency risk, we use the absolute value of the estimated exposure coefficient as dependent variable and estimate δ7 and δ8 to measure the effectiveness of operational and financial hedges in reducing the firm’s absolute currency exposure. SIZEi is measured by the firms’ end of year total assets (in millions of dollars). LTD/TAi is long-term debt to total assets. INSDPi is percentage of common stock outstanding held by insiders; INSDPSQi is INSDPi squared. BLOCPi is the percentage of stock held by blockholders. INSTPi is the percentage of stock held by institutions; MNCi is a proxy for operational hedging, and DER/FSi is the total notional value of currency derivatives divided by foreign sales. Table 1 provides a summary of the variables used in this study as well as the expected relationships between these variables and the absolute value of the estimated foreign exchange coefficient. [Insert Table 1 about here.] Size is included in the model as a proxy for the firm’s expertise in devising hedges (see Booth, Smith, and Stolz (1984)). Thus, we expect a negative relationship between size and exposure. Géczy, Minton, and Schrand (1997) provide evidence that firms hedge to reduce the costs of financial distress. We include the long-term debt ratio (LTD/TA) in our model in order to control for the impact of long-term leverage on firms’ exposures. If LTD/TA proxies for expected bankruptcy costs then the relationship between long-term debt ratio and exposure should be negative. This occurs because bondholders may encourage managers of financially distressed firms to hedge in order to reduce cash flow volatility. In addition, LTD/TA may include a foreign-debt component, which provides a natural hedge against foreign currency exposure.9
This would also imply a
negative relationship between LTD/TA and the absolute value of the exposure coefficient. Whidbee and Wohar (1999) document that managerial incentives and external monitoring 8 Jorion (1990) notes that two-step estimation procedures of this type can possibly produce biased standard errors in the second-stage regression. This occurs if errors in (1) are correlated across stocks and exposure coefficients estimated over the same sample period, yielding estimated exposure coefficients that are not independent across equations. However, Jorion (1990) reports that most of the correlation is due to stock market movements, and is purged by the inclusion of the market return in our equation (1).
8
affect the decision to use derivatives. Also, as shown in Doukas, Hall and Lang (1999), if currency risk is priced, then the shareholders’ concern regarding currency-risk exposure, coupled with managers’ risk aversion can explain why firms may elect to implement extensive foreign-exchange risk-management programs.
We include four equity ownership variables (INSDPi, INSDPSQi,
BLOCPi, and INSTPi) in our cross-sectional model to account for the likelihood that the firm’s foreign-exchange exposure may be affected by its ownership structure. Smith and Stulz (1985) point out that managers who are over-invested in their own firms are more likely to hedge risk exposures, because that reduces the uncertainty associated with their own wealth and decreases the likelihood that they are disciplined for poor firm performance. This is supported by evidence from the goldmining industry [Tufano (1996)].
However, Géczy, Minton, and Schrand (1997) fail to find a
significant relationship between currency risk-hedging and managerial incentives. Smith and Stulz (1985) and Froot, Scharfstein, and Stein (1993) argue that value-maximizing firms hedge risk exposures in order to reduce costs associated with agency conflicts between equity- and debtholders (in particular, the under-investment problem). These studies imply that managerial hedging decisions may be value increasing or decreasing, depending on the managerial incentives. The alignment of managerial incentives with those of the shareholders or the lack thereof will affect the hedging behavior of firms. For low managerial ownership levels, we hypothesize that, as managerial ownership increases managers will engage in less excessive hedging. However, when managers own many shares (i.e. when they are entrenched), they would have more incentives to (excessively) hedge in order to reduce their own personal wealth portfolio’s risk (since they are overinvested in their own firm), potentially resulting in less exposure for such firms. Therefore, to capture this non-linear effect, we include both INSDPi, (hypothesize a positive coefficient) and INSDPSQi (hypothesize a negative coefficient) in the regressions.10
9
Kedia and Mozumbar (1999) show that the fraction of debt denominated in foreign currency is strongly related to the degree of foreign exposure. 10 Other studies have used insider-squared variables to account for curvilinear managerial ownership effects. For example, McConnell and Servaes (1990, 1995) find that market value first increases and then decreases with managerial shareholdings. They found that for low levels of managerial ownership there is a convergence of interests between managers and shareholders, while beyond a critical value (usually around 50% ownership –
9
Outsiders, who have the incentive to oversee the manager’s activities because they want to protect their own investment in the firm, can monitor hedging decisions. This is particularly the case for outsiders who own a substantial fraction of the firm’s equity. Blockholders (owners of 5% or more of the common shares outstanding) will monitor managerial hedging decisions and tend to favor risk management activities that reduce exposure. This is because blockholders, like the entrenched managers, are likely to have a disproportionately large fraction of their wealth invested in the firm. Thus, we expect a negative relationship to exist between block-holdings (BLOCPi) and risk-exposure levels. In addition to the effect of blockholders, we consider the impact of institutional shareholders (INSTPi) on exposure.
Although institutional shareholders may not be blockholders, they are
equipped with the resources to analyze managerial actions. However, the ability of institutional shareholders to diversify their portfolios may result in less monitoring intensity with respect to managerial hedging decisions.
Therefore, the expected relationship between INSTPi and the
absolute value of the exposure coefficient is ambiguous. Alternatively, institutional stakes can be viewed as a proxy for short-selling constraints (see, Nagel (2003), among others). Then, stocks of firms with low (high) institutional shareholdings would be harder (easier) to sell short.11 Since demand for shorts increases with stock price level (Figlewski (1981)) and volatility (Boehme et al (2002)) managers may decide to hedge to reduce that demand and the looming subsequent price correction. In other words, if hedging lowers volatility for the stock price and the demand for shortselling, then we would expect firms with lower institutional shareholdings to hedge in order to alleviate the price pressure that arises from the short-sale constraints. We also investigate the effect of different types of currency derivatives on foreign exchange exposure. We accomplish this by replacing the DER/FSi variable that aggregates all types of currency derivatives with four different variables measuring the usage of options, forwards, swaps, and other currency derivatives. This relationship is analyzed in Equation (3) below:
consistent with the model by Stulz (1988)), there is a reversal, due to entrenched managers making decisions that may not be value increasing.
10
⎮γ i⎮ = δ0 + δ1SIZEi + δ2LTD/TAi + δ3INSDPi + δ4INSDPSQi + δ5BLOCPi + δ6INSTPi + δ7MNCi + (3) δ8OPT/FSi + δ9 FWD/FS i + δ10SWAP/FSi + δ11OTH/FSi + εi, where OPT/FSi is measured as the total notional principal amount of currency options divided by foreign sales; FWD/FS i is the total notional principal amount of currency forwards divided by foreign sales; SWAP/FSi is the total notional principal amount of currency swaps divided by foreign sales; and OTH/FSi is the total of other currency derivatives (such as futures, caps, floors, forward rate agreements, etc.) divided by foreign sales. When analyzing both Equations (2) and (3), we employ weighed least squares and weight our independent variables by the reciprocal of the squared standard error of the exposure coefficient from estimating equation (1).
This procedure gives more weight to the exposure
coefficients that are estimated more precisely.
IV.
Data Sources and Sample Selection Data utilized in this study are obtained from a variety of sources. Firm-specific financial
data used in the cross-sectional regressions are obtained from COMPUSTAT for the year 1996, the mid-point of our study. We collect the 1996 common equity ownership data from Disclosure. Data on MNCs’ financial hedging practices are obtained from the SWAPS Monitor database for 1996. We select MNCs from the cross-section of NYSE-, AMEX- and NASDAQ-listed U.S.-based mining and manufacturing firms (i.e., firms with SICs of 3999 or less). A MNC is defined as a firm that has at least one majority owned foreign subsidiary using information collected from Dunn and Bradstreet’s Who Owns Whom. This results in an initial sample of 340 firms. A final sample size of 208 U.S.-based MNCs, is obtained for all MNCs for which complete financial, ownership, and subsidiary location data are available. Security prices are obtained from the University of Chicago CRSP database for the sample period January 1994 to December 1998. Two different measures of the market return, Rmt,, are used: (1) the value-weighted CRSP market portfolio return (VWR) and (2) the equally-weighted CRSP market portfolio return (EWR). The VWR is commonly used in prior studies on exchange
11 Chen, Hong and Stein (2003) show that low breadth of ownership (i.e., low institutional holdings) indicates that few investors have long positions in the stock, which signals that the short-selling constraint is binding, implying high prices relative to the fundamentals and low expected returns.
11
rate exposure, so it is used here for consistency. However, Bodnar and Wong (2003) point out that the VWR can distort the sign and size of the resulting exposures because of an inherent relation between market capitalization and exposure. They state that exposure studies using the VWR control not only remove the “macroeconomic” effects from the exposure estimates, but also cause a distribution shift in the positive direction for exposure estimates.
They show that
replacing the VWR with the EWR in Equation (1) results in residual exposures that are more consistent with the actual cash flow impact of exchange rate changes predicted by corporate finance models. For this reason, Bodnar and Wong recommend using the EWR control to correct for the potential bias of the VWR control and to achieve less distortion in the residual exposures. Changes in international trading relationships makes it necessary to select exchange rate indexes comprised of currencies which most accurately reflect U.S. competitiveness in world markets. Newer indexes created by the Federal Reserve Board (FRB) focus more directly on the dollar’s movements in foreign exchange markets against a broader set of currencies and are designed to reflect the changing structure of trade patterns and exchange rates.
The FRB
updates currency weights in their indexes annually to most accurately measure the U.S. dollar’s competitive value. For these reasons, we select indexes published by the FRB to compute monthly changes in the value of the dollar, Rct.12 Appendix 1 lists these indexes together with the currencies and weights. These indexes are: (1) the nominal Broad Index, which consists of thirty-five currencies of all foreign countries or regions that had a share of U.S. non-oil imports or nonagricultural exports of at least 0.5 percent in 1997, (2) the nominal Major Index, which includes sixteen currencies traded in deep and relatively liquid financial markets and for which short- and long-term interest rates are readily available, (3) the nominal OITP (Other Important Trading Partners) Index, which includes nineteen currencies of key U.S. trading partners in Latin America, Asia, the Middle East, and Eastern Europe, and (4) the real Broad Index, which captures the real (price adjusted) exchange rate value of the Broad Index. The Major and OITP indexes are subsets of the Broad Index.
12 For more information on the currency indexes, refer to Leahy (1998) or go to the Federal Reserve Board website at www.federalreserve.gov.
12
The exchange rate exposure of the sample firms is estimated from Equation (1) using the two different market controls indices (VWR and EWR) with each of the four foreign exchange indexes, resulting in a total of eight different estimates of foreign exchange exposure. This allows us to examine the robustness of the results using the most commonly used market portfolio control (VWR) with results obtained using the EWR market portfolio control recommended by Bodnar and Wong (2003). Table 2 provides summary descriptive statistics for the exposure coefficients obtained using the eight different combinations of market and foreign exchange indices. As shown, the results are generally consistent regardless of the way the exposure coefficients are computed. All mean exposures are negative implying that on average, MNCs benefit when the dollar depreciates and vice versa. For example, the mean exposure coefficient for the nominal Broad Index using the value-weighted market control (NBVWR) is –0.5274. This indicates that, on average, a one percent increase in the dollar’s value results in a 0.53 percent decline in MNC stock prices. Approximately 30 to 40 percent of MNCs are positively exposed to the dollar’s value and the range of exposure coefficients are largest when using the broad index (both nominal and real index values). For brevity, the remaining tables in the paper report results using the Broad Index (both nominal and real values) only. It should be noted that our results are robust in all our tests to the eight alternative methods of estimating foreign exchange exposure. [Insert Table 2 about here] Table 3 describes the use of currency derivatives by sample firms. As shown in Panel A, 119 MNCs (57.2 percent of the sample) use currency derivatives. The mean level of currency derivatives by sample firms is $0.213 notional principal amount per dollar of foreign sales with the highest and lowest levels being 1.685 and 0.0005, respectively. The highest mean level is for currency forwards ($0.12 notional principal value per dollar foreign sales), followed by other currency derivatives ($0.052 notional principal value per dollar foreign sales). As shown in Panel B, the most frequently used contracts are currency forwards (used by 44.2 percent of the MNCs), followed by currency swaps (used by 14.9 percent of the MNCs) and currency options (used by 10 percent of the MNCs). It is interesting to note, that regardless of the
13
currency exposure measure used, exposure is greater for firms that do not use currency derivatives (with the exception of options users with exposure coefficients estimated using the real Broad index with the equally-weighted CRSP market index). This result is consistent with our hypothesis that currency derivatives are used to reduce the foreign exchange exposure of MNCs.13 We find statistically significant differences in mean absolute exposure coefficients for users of currency swaps and users of other foreign-exchange derivatives as compared to nonusers of derivatives. These differences are significant at the one-percent level, or better. [Insert Table 3 about here] A MNC’s operational hedging strategy is difficult to observe and measure. The ability to effectively design and implement operational hedges to reduce operating exposures is expected to be a function of a MNC’s operating flexibility. As shown by Pantzalis, Simkins, and Laux (2001), MNCs with greater breadth of operations are better equipped to effectively manage their foreign-exchange exposure. Alternatively, MNCs with networks that are concentrated in a few countries and/or regions should be less capable of effectively reducing their currency exposures. In our analysis, we use four measures to describe the MNC network of foreign operations (i.e., the breadth and depth dimensions of the network).14 Our four measures are: (1) the number of foreign countries in which the MNC has a subsidiary (NFC); (2) number of foreign regions in which the MNC operates (NFR); (3) a Herfindahl index (HERF1) measuring the concentration of an MNC network’s foreign subsidiaries across different countries; and (4) a Herfindahl index (HERF2) measuring the concentration of an MNC network’s foreign countries across different geographic regions. HERF1 and HERF2 are calculated as follows: HERF1 = Σi (NFSi)2/[ Σi (NFSi)]2 and
(4a)
HERF2 = Σj (NFCj)2/[Σ j (NFCj)]2,
(4b)
where NFSi is the number of foreign subsidiaries in country i and NFCj is the number of foreign 13
In one case the exposure coefficient is higher for users. This occurs for non-users versus options users for the RBEWR exposure coefficient.
14
Breadth describes dispersion of a firm’s operations and is measured by the number of foreign nations in which the MNC has subsidiaries. Depth measures the geographic concentration of an MNCs foreign subsidiaries. Allen and Pantzalis (1996), Doukas, Pantzalis and Kim (1999), and Pantzalis, Simkins, and Laux (2001) demonstrate the importance of breadth and depth measures in describing a MNC’s network.
14
countries in geographic region j. A higher value for either index, HERF1 or HERF2, indicates a more highly concentrated network structure and is hypothesized to be associated with increased foreign exchange exposure.15 A more geographically dispersed operational network (i.e. lower HERF1 or HERF2 index value) is a proxy for high levels of effective operational hedges and, thus, it is hypothesized to be associated with decreased foreign exchange exposure. The information related to the number of subsidiaries and their geographic location was extracted from Dunn and Bradstreet’s Who Owns Whom. This source provides a list of each firm’s domestic and foreign affiliates, and their location. These data were used in the calculation of our four measures of MNC network dispersion. The foreign subsidiaries were assigned to the following nine major geographic regions, depending on the location of each subsidiary’s host country. The nine different regions are constructed as follows: 1) The East Asia region consists of the following countries: Brunei, China, Hong-Kong, Indonesia, Japan, Korea, Macao, Malaysia, Papua-New Guinea, Singapore, Taiwan, Thailand, Vietnam, and the Philippines. 2) The NAFTA region consists of the following countries: Canada and Mexico. The U.S. is of course part of NAFTA, but U.S. subsidiaries are not counted as part of the NAFTA operations of the sample’s MNCs because they are domestic rather than foreign subsidiaries of our sample’s U.S.-based MNCs. 3) The European Union region consists of the following EU member-countries: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxemburg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom. 4) The Other Asia region includes all Asian countries not accounted for in the East Asian region. 5) The Western Europe region includes non-EU Western European countries and some Eastern European countries that have recently applied for EU membership and are expected to join the EU in the next expansion round. They are the following: Norway, Switzerland, Cyprus, Hungary, Slovenia, Poland, Czech Republic, and Malta. 6) The Eastern Europe region includes all remaining former Soviet Block and Socialist countries. 7) The Central America and Caribbean region includes all Caribbean and Central American nations.
15 For example, assume a MNC has three subsidiaries located in three different countries, HERF1 would be 2 2 2 2 0.333 [i.e. (1 + 1 + 1 )/ 3 ) = 0.333]. If all three subsidiaries are located in the same country, HERF1 is 1.0 2 2 [i.e. 3 /3 = 1.0]. HERF2 is calculated similarly. Note that the highest value for HERF1 or HERF2 is 1.0, indicating complete concentration in a single country or region.
15
8) The South America region includes all nations located in the South American continent. 9) The Africa region includes all countries of the African continent. The following section provides descriptive statistics for our operational hedges and presents cross-sectional results for examining the hypothesized relations between financial and operational hedges and currency exposure. V.
Empirical Results
A.
Descriptive Statistics Table 4, Panel A, reports descriptive statistics for four measures of foreign network
dispersion for our sample of U.S. MNCs. These measures are used as proxies of the MNC’s ability to devise operational hedges. We argue that MNCs with networks of operations that span many different countries and/or regions will be in a better position to adjust operations to effectively hedge their operating exposure. We use the variables NFC, NFR, HERF1, and HERF2, discussed in the previous section, to describe the dispersion/concentration of a MNCs foreign network. Note that higher values for NFC (the number of foreign countries in which it has operations) and NFR (the number of foreign regions in which it has operations) denote a more dispersed network. On the other hand, high values of the two alternative Herfindahl indices indicate a greater degree of concentration of a firm’s operations in a particular country (HERF1) or region (HERF2). As shown in Panel A, there are wide variations in the number of countries (NFC), number of regions (NFR), and the two Herfindahl ratios across sample firms. The average MNC in our sample operates in approximately thirteen foreign countries and four different geographic regions. The minimum and maximum number of foreign countries (foreign regions) are 1 and 47 countries (1 and 9 regions), respectively. The average Herfindahl index ranges from 0.237 (HERF1) to 0.448 (HERF2). [Insert Table 4 about here] Panel B of Table 4 compares the means of the absolute values of the estimated exposure coefficients for firms with low and high values of the foreign operations variables (e.g.,
16
NFC, NFR, HERF1, or HERF2).16 The t-tests for differences in the mean exposure for firms with high levels of dispersion in their foreign operations versus those with low dispersion suggest statistically significant differences. With the exception of the results for HERF1, which are not statistically significant, we find that firms operating in only a few countries or regions, or are highly concentrated in a particular region, have significantly higher levels of foreign-exchange exposure. This result is consistent with our earlier hypothesis that firms operating in a large number of foreign countries or regions are less exposed to foreign-exchange risk. B.
Regression Results Results of our regression analyses for Equations (2) and (3) are presented in Table 5.
Panels A and B report the determinants of absolute exposure coefficients for the dependent variables |NBEWR| (absolute value of the foreign-exchange coefficient estimated using the nominal Broad Index with EWR control) and |RBEWR| (absolute value of the foreign-exchange coefficient estimated using the real Broad Index with EWR control). In each panel, columns 1 through 4 present findings for the operational hedging measures NFC, NFR, HERF1, and HERF2, and sub-columns a and b of each column contain results for the models presented in equations (2) and (3), respectively.17 [Insert Table 5 about here] The estimated coefficients for derivatives usage (DER/FS), in our various models (1a, 2a, 3a, and 4a), range from –0.7903 to –0.9857 and are all significant at the one-percent level or better. These results suggest that the use of currency derivatives is associated with a reduction in foreign-exchange exposure. This implies that MNCs are using currency derivatives to reduce their exposure to unexpected changes in foreign-exchange rates. In addition to examining the effect of currency derivatives in the aggregate, we also examine the impact of different types of derivatives on the foreign-exchange exposure of our sample firms. These results are found in models 1b, 2b, 3b, and 4b. With the exception of currency options (OPT/FS), all other foreign-
16 A firm is assigned to the low (high) group if the value of its foreign operations variable (e.g., NFC, NFR, HERF1, or HERF2) is below (above) the sample median for that variable. 17
While not reported, the results in which the dependent variable was estimated using the CRSP valueweighted market index (NBVWR and RBVWR) are similar to those presented in Table 5.
17
exchange derivatives have negative coefficients, although the estimates for swaps (SWAP/FS) are not statistically significant. The estimated coefficients for options are generally positive but not significant in any of the models.
Both currency forwards (FWD/FS) and other foreign-
exchange derivatives (OTH/FS) have negative and statistically significant estimates. Depending on the model, the estimates for currency forwards range from –1.2053 to –1.4950 and those for other foreign-exchange derivatives range from –0.7723 to –0.9793. The estimates for these two variables are statistically significant at better than the one-percent level in all cases. These results show that the use of currency forwards and other derivatives have the effect of decreasing the firm’s foreign-exchange exposure. Neither foreign-exchange options nor swaps appear to be associated with a reduction, or an increase, in exposure. The results for the MNC network support our hypothesis that firms benefit from spreading their operations across many countries or geographic regions. The estimated coefficients for measures of the number of countries (NFC) or regions (NFR) in which a firm operates are negative and significant in all of our models. For example, the estimated coefficients for NFC (Panel A, models 1a and 2a, and Panel B, models 1a and 2a) range from –0.0373 to –0.0445 and are significant at better than the five-percent level in all estimations (the estimates are actually significant at better than the one-percent level when the dependent variable is |RBEWR|). Alternatively, our measures of concentration (e.g., HERF1 and HERF2) have significant positive estimates. Thus, firms operating in a larger number of countries or regions tend to have lower foreign-exchange exposures, while those that are concentrated in a few countries or regions tend to have higher exposures. These results are robust to various specifications of the hedging variable and currency risk exposure measures, and highlight the importance of MNC network structure in enabling the firm to develop operational hedges. We include several ownership/corporate control variables in our regressions.
These
variables are: percentage of ownership by insiders (INSDP), the square of INSDP (INSDPSQ), the percentage of ownership by blockholders (BLOCP), and the percentage of ownership by institutions (INSTP). Of these ownership/control variables, only institutional ownership (INSTP) is significant. In all models, the coefficient is positive and significant at the one percent level or
18
better. This finding is consistent with the notion that institutional investors are less concerned with foreign exchange risk, and therefore they are willing to tolerate higher exposures. Alternatively, this finding suggests that in the absence of significant institutional holdings managers have an incentive to hedge to reduce potential price pressure effects related to demand for short-selling. The estimated coefficients for leverage (LTD/TA) are negative and significant at the ten percent level or better (the estimates shown in Panel A, models 3a and 3b, and Panel B, models 1a, 1b, 3a, and 3b are significant at the five percent level or better). There are three likely explanations for this result: first, firms with higher levels of debt in their capital structures would have a higher probability of financial distress. Thus, value-maximizing managers would try to hedge to decrease the volatility of cash flows, which would reduce exposure. Second, higher monitoring intensity by debtholders, who have an incentive to reduce cash flow volatility, may cause managers to engage in hedging activities and thus reduce exposure.
Third, firms
sometimes resort to foreign-currency-denominated debt as a means of hedging. If the greater use of foreign-currency-denominated long-term debt results in greater long-term debt to assets ratios, then the negative relationship between LTD/TA and exposure could be attributed to hedging via foreign currency long-term debt financing. To summarize, we find strong evidence that the number of countries or regions, in which a firm operates, is associated with lower absolute foreign-exchange exposure.
Further,
concentration in a few countries or regions increases exposure to changes in foreign-exchange rates. We also find the use of currency derivatives, particularly forwards and other derivatives, is useful in reducing exposure. These results are robust to a variety of specifications and measures of exposure. C.
Firmwide Risk Management The regression results we present in Table 5 are consistent with the hypothesized
relationships between foreign-exchange exposure and the existence of financial and operating hedges. Next, we turn to the investigation of the coordinated use of these hedges to reduce exposure (i.e., firmwide risk management).
19
We create two variables, FIMWIDE1 and FIRMWIDE2, to explore the relationship between foreign-exchange rate exposure and coordinated (i.e., firmwide) risk management practices by MNCs. The variable FIRMWIDE1 equals zero if the number of countries in which the firm has operations (NFC) is less than the sample median and DER/FS is less than the median. Alternatively, FIRMWIDE1 equals one if either NFC or DER/FS is greater than the median. Our other measure of firmwide risk management, FIRMWIDE2, is similar except it is coded as either one or zero using the number of regions in which the firm operates (NFR). Firms operating in a small number of countries or regions and making little use of derivatives will be coded as a zero. Those MNCs operating in a larger number of countries or regions and/or using derivatives to a greater extent will be coded as a one. Thus, our firmwide measures capture coordinated riskmanagement practices by MNCs. [Insert Table 6 about here] The results for our regressions using the firmwide measures are reported in Table 6. We present results using both measures of foreign-exchange exposure (|NBEWR| and |RBEWR|) and both firmwide variables. Additionally, we include DER/FS in some of the models to see if the significant negative relationship between exposure and currency derivatives usage, reported in Table 5, persists after controlling for firmwide risk-management practices. Together, we report the results of eight separate regression models. Models 1a and 1c (2a and 2c) present the results for the NBEWR (RBEWR) exposure coefficients and firmwide measures FIRMWIDE1 and FIRMWIDE2. In all cases, the parameter estimates for the firmwide variables are negative and statistically significant at the one-percent level or better. These results strongly support the notion that MNCs using currency derivatives and operating hedges in an integrated manner can significantly reduce exposure to currency risk. Models 1b and 1d (2b and 2d) report the results for regression in which the variables HERF1 and DER/FS are also added to the model. In each of these cases, the firmwide measure is negative and significant at the ten-percent level or better. The coefficient estimates for the derivatives use and the MNC network concentration variables have the expected negative and positive sign, respectively. However, the coefficients of DER/FS and HERF1 are not significant,
20
suggesting that there is no further explanatory power gained beyond that provided by considering coordinated risk-management practices by the firm.
These results strongly support the
importance of integrated or firmwide risk-management programs by MNCs.
V.
Conclusion This study investigates the influence of financial and operating hedges on the foreign-
exchange exposure of U.S. multinational corporations. We build on previous studies of currency exposure to more fully understand how corporate risk management practices can reduce exposure to exchange-rate risk by managing risk across the firm. Our research is important for the following two reasons. First, the evidence indicates that operational hedges and financial hedges can effectively reduce foreign currency exposure. The ability to construct operational hedges, reflected in the MNC network structure, and the usage of various currency derivatives are significant determinants of currency exposure.
Furthermore, these results are robust to
alternative methods of measuring currency exposure, operating hedges, and financial hedges. Second, our results indicate that multinational corporations are taking a firmwide or “strategic” perspective in their currency risk management strategy, and are thus focusing on hedging overall economic exposure. We present strong evidence that shows the combined use of operating hedges and financial hedges is associated with decreased exchange-rate exposure. If firms were speculating using currency derivatives, or not attempting to use operational hedges effectively, we should find financial hedges and operational hedges associated with increased exposure. However, this is not the case. These results are consistent with the notion that operational and financial hedges compliment each other and support the importance of firmwide risk management in mitigating currency risk.
21
References Allen, Linda and Christos Pantzalis, 1996, Valuation of the operating flexibility of multinational corporations. Journal of International Business Studies 27: 633-653. Allayannis, George, 1996, The time-variation of the exchange rate exposure: An industry analysis, Working Paper, University of Virginia. Allayannis, George and E. Ofek, 2001, Exchange rate exposure, hedging, and the use of Foreign Currency Derivatives, Journal of International Money and Finance 20, 273-296. Allayannis, George, Jane Ihrig and James P. Weston, 2001, Exchange rate hedging: Financial vs. operational strategies, American Economic Review Papers and Proceedings 91 (2): 391-395. Banham, Russ, 2000, Ford’s better idea, Treasury and Risk Management, September 2000. Bartov, Eli and Gordon M. Bodnar, 1994, Firm valuation, earnings expectations, and the exchange rate exposure effect, The Journal of Finance 44: 1755-1785. Bartov, E., G.M. Bodnar, and A. Kaul, 1996, Exchange rate variability and the riskiness of U.S. multinational firms: Evidence from the breakdown of the Bretton Woods system, Journal of Financial Economics 42: 105-132. Belsley, D.A., E. Kuh and R.E. Welsch, 1980, Regression Diagnostics: Identifying Influential Data and Sources of Collinearity, John Wiley & Sons, Inc., New York. Bodnar, Gordon M. and W.M. Gentry, 1993, Exchange rate exposure and industry characteristics: Evidence from Canada, Japan and U.S., Journal of International Money and Finance 12: 2945. Bodnar, Gordon M. and M.H. Franco Wong, 2003, Estimating exchange rate exposures: Issues in model structure, Financial Management 32: 35-67. Bodnar, Gordon M., Gregory S. Hayt, and Richard C. Marston, 1998, 1998 Wharton survey of financial risk management by U.S.non-financial firms, Financial Management 27: 70-91. Bodnar, Gordon M., Hayt, G.S., Marston, R.C. and C.W. Smithson, 1995, Wharton survey of derivatives usage by U.S. non-financial firms, Financial Management 24: 104-114. Boehme, Rodney D., Brtley R. Danielsen and Sorin M. Sorescu, 2002, Short-sale constraints and overvaluation, University of Houston Working Paper. Booth, J.R., Smith, R.L. and Stolz, R.W., 1984, Use of interest rate futures by financial institutions, Journal of Bank Research 15: 15-20. Breeden, Douglas and S. Viswanathan, 1996, Why do firms hedge: An asymmetric information model, Fuqua School Working Paper. Buckley, Peter J. and Mark C. Casson, 1998, Models of the multinational enterprise, Journal of International Business Studies 29: 21-44. Chen, Joseph, Harrison Hong and Jeremy C. Stein, 2003, Breadth of ownership and stock returns, Journal of Financial Economics 66: 171-205. Christophe, Stephen E., 1997, Hysteresis and the value of the U.S. multinational corporation, Journal of Business 70: 435-462.
22
Choi, Jongmoo J. and Anita M. Prasad, 1995, Exchange risk sensitivity and its determinants: A firm and industry analysis of U.S. multinationals, Financial Management 24: 77-88. Chow, Edward H., Lee, Wayne Y., and Michael E. Solt, 1997a, The economic exposure of U.S. multinational firms, The Journal of Financial Research 20: 191-210. Chow, Edward H., Lee, Wayne Y., and Michael E. Solt, 1997b, The exchange-rate risk exposure of asset returns, Journal of Business 70: 105-123. Chowdhry, Bhagwan and Jonathan Howe, 1999, Corporate risk management for multinational corporations: Financial and operating hedging policies, European Financial Review 2: 229246. Cox, W. Michael, 1986, A new alternative trade-weighted dollar exchange rate index, Federal Reserve Bank of Dallas, Economic Review 67: 20-28. Davis, Henry A. and Frederick C. Militello, Jr., 1995, Foreign Exchange Risk Management: A Survey of Corporate Practices, Financial Executives Research Foundation, Morristown, New Jersey. Deragon, Jay, 2000, Old knowledge with a new name, Erisk.com, November 2000. Directory of International Affiliations, 1992, National Register Publishing Co., Skokie, Illinois. Dolde, Walter, 1995, Hedging, leverage, and primitive risk, The Journal of Financial Engineering 4: 187-216. Doukas, John, Patricia H. Hall and Larry H.P. Lang, 1999, The pricing of currency risk in Japan, Journal of Banking and Finance 23: 1-20. Doukas, John, Christos Pantzalis, and Sungsoo Kim, 1999, Intangible assets and the network structure of MNCs, Journal of International Financial Management and Accounting 10: 1-23. Dumas, Bernard and Bruno Solnik, 1995, The world price of exchange rate risk, Journal of Finance 50: 445-479. Errunza, Vihang R. and Lemma W. Senbet, 1981, The effects of international operations on the market value of the firm: Theory and evidence, Journal of Finance 36: 401-417. Errunza, Vihang R. and Lemma W. Senbet, 1984, International corporate diversification, market value, and size-adjusted evidence, Journal of Finance 39: 727-743. Figlewski , Stephen, 1981., The informational effects of restrictions on short sales: Some empirical evidence, Journal of Financial and Quantitative Analysis 16: 463-476. Flood, Eugene and Donald R. Lessard, 1986, On the measurement of operating exposure to exchange rates: A conceptual approach, Financial Management 15: 25-36. Froot, Kenneth, 1993, Currency hedging over long horizons, NBER Working Paper. Gay, Gerald D. and Jouahn Nam, 1998, The underinvestment problem and corporate derivatives use, Georgia State University Working Paper. Géczy, Christopher, Bernadette A. Minton, and Catherine Schrand, 1997, Why firms use currency derivatives, The Journal of Finance 52: 1323-1354.
23
Greene, William H., 2000, Econometric Analysis, Prentice Hall: Upper Saddle River, NJ. Jensen, Michael C., 1986, Agency costs of free cash flow, corporate finance and takeovers, American Economic Review 76: 323-329. Jensen, Michael C. and Kevin J. Murphy, 1990, CEO incentives- it’s not how much you will pay, but how, Journal of Applied Corporate Finance 3: 36-49. Jesswein, Kurt, Chuck C.Y. Kwok and William R. Folks, Jr., 1995, What new currency risk products are companies using and why?, Journal of Applied Corporate Finance 8: 103-114. Jorion, Philippe, 1990, The exchange rate exposure of U.S. multinationals, Journal of Business 63: 331-45. Jorion, Philippe, 1991, The pricing of exchange rate risk in the stock market, Journal of Financial and Quantitative Analysis 26: 363-376. Kedia, Simi, and Abon Mozumbar, 1999, Is foreign currency denominated debt a hedging instrument? Working paper Harvard University and Virginia Tech University. Kim, Wi Saeng and Esmeralda O. Lyn, 1986, Excess market value, the multinational corporation, and Tobin’s q-Ratio, Journal of International Business Studies 17: 119-125. Kogut, Bruce, 1983, Foreign direct investment as a sequential process, in C.P.Kindleberger and D. Audresch, editors, The Multinational Corporation in the 1980’s, Cambridge, Mass.: MIT Press. Kogut, Bruce, and Nalin Kulatilaka, 1995, Operating flexibility, global manufacturing, and the option value of a multinational network, Management Science 40: 123-39. Leahy, Michael P., 1998, New summary measures of the foreign exchange value of the dollar, Federal Reserve Bulletin, October, 811-818. Lehtinen, Risto, 1996, Firmwide cashflows, Firmwide Risk Management: A Risk Special Supplement, July, 30-31. Logue, Dennis, 1995, When theory fails: Globalization as a response to the (Hostile) Market for Foreign Exchange, Journal of Applied Corporate Finance 8: 39-48. McConnell, John J. and Henri Servaes, 1990, Additional evidence on equity ownership and corporate value, Journal of Financial Economics 27: 595-612. McConnell, John J. and Henri Servaes, 1995, Equity ownership and the two faces of debt, Journal of Financial Economics 39: 131-157. Mello, Antonio S., John E. Parsons, and Alexander Triantis, 1995, An integrated model of multinational flexibility and hedging policies, Journal of International Economics 33: 41-56. Mian, Shehzad L., 1996, Evidence on corporate hedging policy, Journal of Financial and Quantitative Analysis 31: 419-39. Modigliani and Miller, 1958, The cost of capital, corporation finance, and the theory of investment, American Economic Review 48: 261-297.
24
Montgomery, Douglas C. and Elizabeth A. Peck, 1992, Introduction to Linear Regression Analysis, 2nd edition, John Wiley & Sons, Inc., New York. Morck, Randall and Bernard Yeung, 1991, Why investors value multinationality, Journal of Business 64: 165-187. Nagel, Stefan, 2003, Short sales, institutitonal investors, and the book-to-market effect, London Business School Working paper. Nance, Deana R., Clifford W. Smith, Jr., and Charles W. Smithson, 1993, On the determinants of corporate hedging, Journal of Finance 48: 267-284. Pantzalis, Christos, Betty J. Simkins, and Paul Laux, 2001, Operational hedges and the foreign exchange exposure of U.S. multinational corporations, Journal of International Business Studies. Paul-Choudhury, Sumit, 1996, “Crossing the Divide”, Firmwide Risk Management, A Special Supplement to Risk July, Risk Publications, 2-4. Phillips, Aaron L., 1995, 1995 Derivatives practices and instruments survey, Financial Management 24: 115-125. Shapiro, Alan C., 1996, Multinational Financial Management, 5th edition, Prentice Hall Inc., Upper Saddle River, New Jersey. Simkins, Betty J. and Paul Laux, 1996, Derivative use and the exchange rate risk of large U.S. corporations, Conference Proceedings: The Second International Finance Conference, Georgia Tech University. Stulz, Rene M., 1988, Managerial control of voting rights: Financing policies and the market for corporate control, Journal of Financial Economics 20: 25-54. Stulz, Rene M., 1990, Managerial discretion and optimal financing policies, Journal of Financial Economics 26: 3-28. Stulz, René M., 1996, Rethinking risk management, Journal of Applied Corporate Finance 9: 824. Whidbee, David A., and Mark Wohar, 1999, Derivative activities and managerial incentives in the banking industry, Journal of Corporate Finance 5: 251-276.
25
Table 1 Description of Variables and Hypothesized Relations Between MNC Characteristics and the Absolute Value of the Foreign Exchange Exposure Coefficient This table provides a description of the variables used in the analysis. All variables are calculated for the year 1996. There are nine different geographic regions used in constructing the operational hedging variables: East Asian region, NAFTA, European Union, Other Asia region, Western Europe region, Eastern Europe region, Central America and Caribbean, South America and Africa. Hypothesis Financial Hedges
Operational Hedges
Control Variables: Size
Variable
Sign
Data Description (Source)
DER/FS
-
OPT/FS
-
FWD/FS
-
SWAP/FS
-
OTH/FS
-
NFC
-
NFR
-
HERF1
+
HERF2
+
SIZE
-
LTD/TA
-
INSDP
+
Calculated as the total notional principal amount of currency derivatives/total foreign sales. This variable includes currency derivatives such as forwards, futures, swaps, caps, collars and others. (Swaps Monitor Database) Calculated as the total notional principal amount of currency options/total foreign sales. (Swaps Monitor Database) Calculated as the total notional principal amount of currency forward contracts/total foreign sales. (Swaps Monitor Database) Calculated as the total notional principal amount of currency swaps/total foreign sales. (Swaps Monitor Database) Calculated as the total notional principal amount of currency derivatives other than options, forwards, and swaps. (i.e. includes caps, collars, floors, swaptions, etc) (Swaps Monitor Database) Number of foreign countries in which the MNC has a subsidiary. (Who Owns Whom) Number of foreign regions in which the MNC operates. (Who Owns Whom) Herfindahl Index 1. The index is calculated as follows: 2 2 HERF1 = Σ i (NFSi) /[ Σi (NFSi)] , where NFSi is the number of foreign subsidiaries in country i. Herfindahl Index 2: The index is calculated as follows: 2 2 HERF2 = Σj (NFSj) /[Σ j (NFSj)] , where NFSj is the number of foreign subsidiaries in geographic region j. Total Assets (in million dollars); (Compustat) Long-term debt/Total Assets (Compustat) Percentage of stock held by insiders. (Compact Disclosure) The squared value of the insider ownership variable
Financial Distress Insider Ownership INSDPSQ
-
Insider Ownership Squared
BLOCP
-
Blockholders
INSTP
+/-
Institutional Ownership
Percentage of stock held by blockholders. (Compact Disclosure) Percentage of stock held by institutions. (Compact Disclosure)
26
Table 2 Descriptive Statistics of Exposure Coefficients Reported are the minimum, maximum, quartiles, means, standard deviations and ranges for the following eight alternative foreign exchange exposure coefficients: NBEWR is the exposure coefficient obtained using the nominal broad exchange rate index and an equally weighted market portfolio. NBVWR is the exposure coefficient obtained using the nominal broad exchange rate index and a value weighted market portfolio. NMEWR is the exposure coefficient obtained using the nominal major currencies’ exchange rate index and an equally weighted market portfolio. NMVWR is the exposure coefficient obtained using the nominal major currencies’ exchange rate index and a value weighted market portfolio. NOEWR is the exposure coefficient obtained using the nominal other currencies’ exchange rate index and an equally weighted market portfolio. NOVWR is the exposure coefficient obtained using the nominal other currencies’ exchange rate index and a value weighted market portfolio. RBEWR is the exposure coefficient obtained using the real broad exchange rate index and an equally weighted market portfolio. RBVWR is the exposure coefficient obtained using the real broad exchange rate index and a value weighted market portfolio. All exposure coefficients (γi) were obtained from the following time series regression: Rit = αi + βiRmt + γiRct + εit,where Ri,t is the return of stock i in month t, RM,t is the return on the market portfolio in month t, and RC,t is the percent change in the exchange rate index in month t. Regressions were run over a 60-month period covering the years 19941998. [N=208]
NBEWR
NBVWR
NMEWR
NMVWR
NOEWR
NOVWR
RBEWR
RBVWR
min
-5.8799
-6.0626
-4.4696
-4.5892
-3.0903
-3.1192
-6.0193
-6.3199
Q1
-0.8494
-0.9176
-0.7709
-0.7740
-0.4920
-0.5288
-0.8496
-1.0045
median
-0.4335
-0.3907
-0.2884
-0.2781
-0.0967
-0.0957
-0.3452
-0.3989
Q3
0.0906
0.1357
0.1885
0.2213
0.1135
0.1299
0.1611
0.1350
max
2.1961
2.1829
2.2099
1.7572
1.3889
1.2707
2.1578
2.2469
mean
-0.0508
-0.5274
-0.3225
-0.3195
-0.2418
-0.2484
-0.5002
-0.5901
std. dev.
1.1117
1.1620
0.9392
0.9556
0.6575
0.6795
1.1376
1.2028
positive
31.25 %
31.73 %
35.10 %
37.50 %
38.46 %
39.90 %
30.29 %
28.85 %
range
8.076
8.2455
6.6795
6.3464
4.4792
4.3899
8.1771
8.5668
27
Table 3 Currency Derivatives Usage by Sample Firms Panel A: Descriptive statistics of currency derivatives’ notional values as a percent of foreign sales for the subsample of the MNC firms that use currency derivatives. Definitions are as follows: DER/FS = Total currency derivatives notional value / foreign sales, OPT/FS = currency options notional value / foreign sales, FWD/FS = currency forwards notional value / foreign sales, OTH/FS = other currency derivatives notional value / foreign sales. [N=119]
DER/FS
OPT/FS
FWD/FS
SWAP/FS
OTH/FS
min
0.0005
0
0
0
0
Q1
0.0475
0
0.0082
0
0
median
0.1249
0
0.0706
0
0
Q3
0.3129
0.0497
0.1406
0.0049
0
max
1.6851
0.2193
0.8890
1.1088
1.5094
mean
0.2130
0.0142
0.1197
0.0350
0.0517
std. dev.
0.2786
0.0447
0.1641
0.1180
0.1834
Panel B: Mean values of absolute exposures for non-users and users of derivatives. |NBEWR| is the currency exposure coefficient obtained using the nominal Broad Index and the equally-weighted CRSP market portfolio return. |NBVWR| is the currency exposure coefficient obtained using the nominal Broad Index and the value-weighted CRSP market portfolio return. |RBEWR| is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the equally-weighted CRSP market portfolio return. |RBVWR| is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the value-weighted CRSP market portfolio return. Note, *, **, and *** denote that the mean value is significantly different than that of the non-users’ group at the 10%, 5% and 1% significance level, respectively. Non-Users
Users
89
All Users 119
Options Users 21
Forwards Users 92
Swaps Users 31
Other derivatives users 30
42.79 %
57.21 %
10.01%
44.23 %
14.90 %
14.42 %
|NBEWR|
0.9073
0.7388
0.8911
0.7918
0.5158 ***
0.5107 ***
|NBVWR|
0.9690
0.7581
0.8381
0.8077
0.4614 ***
0.5514 ***
|RBEWR|
0.8975
0.7740
0.9122
0.8342
0.5406 ***
0.5388 ***
|RBVWR|
0.9866
0.8201
0.8781
0.8772
0.5085 ***
0.5921 ***
N % of all firms
28
Table 4 Foreign Operations Network of Sample Firms The foreign operations network variables are used as proxies for the firms ability to construct operational hedges for foreign exchange risk management. Descriptive statistics (Panel A) and comparisons of mean exposures of different groups of MNCs classified based on whether their foreign operations network variables are above or below the sample median. The following variables are examined: NFC = number of foreign countries the MNC is operating in. NFR = number of foreign regions the MNC operates in. HERF1 is a concentration (Herfindahl) ratio of the number of 2 2 foreign subsidiaries across foreign countries, defined as HERF1 = Σi (NFSi) /[Σi (NFSi)] , where NFSi is the number of foreign subsidiaries in country i. HERF2 is a concentration (Herfindahl) ratio of the number of foreign countries 2 2 across foreign regions, defined as HERF2 = Σj (NFCj) /[Σj (NFCj)] , where NFCj is the number of foreign countries in geographic region j. There are nine different geographic regions: Asian Crisis region, NAFTA, European Union, Other Asia region, Western Europe region, Eastern Europe region, Central America and Caribbean, South America and Africa. Panel A: Descriptive statistics for variables describing the foreign operations network of MNCs. [N=208]
NFC
NFR
HERF1
HERF2
min
1
1
0.0424
0.1728
Q1
5
2
0.0944
0.2717
median
11
4
0.1481
0.3574
Q3
20
6
0.2658
0.52
max
47
9
1
1
mean
13.35
4.22
0.2369
0.4477
Std. dev
10.43
2.13
0.2359
0.2521
Panel B: Comparison of absolute exposures’ mean values for firms with “low” and “high” values of foreign operations variables. A firm is assigned to the “low” (“high”) group if the value of the particular variable for the firm is below (above) the sample median value. Note, *, **, and *** denote that the mean value for the “High” group is significantly different than that of the “Low” group at the 10%, 5% and 1% significance level, respectively. |NBEWR|, is the currency exposure coefficient obtained using the nominal Broad Index and the equally-weighted CRSP market portfolio return. |NBVWR |, is the currency exposure coefficient obtained using the nominal Broad Index and the value-weighted CRSP market portfolio return. |RBEWR|, is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the equally-weighted CRSP market portfolio return. |RBVWR|, is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the value-weighted CRSP market portfolio return. NFC
NFR
HERF1
HERF2
“Low”
“High”
“Low”
“High”
“Low”
“High”
“Low”
“High”
|NBEWR|
0.9520
0.6585 **
0.9561
0.6629 **
0.7235
0.8983
0.6336
0.9882 ***
|NBVWR|
1.0096
0.6742 ***
1.0246
0.6686 ***
0.7498
0.9469
0.6415
1.0552 ***
|RBEWR|
0.9801
0.6613 **
0.9738
0.6770 **
0.7397
0.9140
0.6554
0.9983 ***
|RBVWR|
1.0693
0.6992 ***
1.0783
0.7008 ***
0.7888
0.9939
0.6756
1.1071 ***
29
Table 5 Relationship Between Absolute Exposure and Financial and Operational Hedges This table reports estimates of the determinants of the absolute exchange rate exposure. Weighted least squares regressions with robust standard errors (White (1980)) are used to control for heteroskedasticity. The weighted least squares procedure weights by the exposure coefficient’s standard error. Significance levels are indicated as follows: *** 1%, ** 5%, * 10%. The independent variables are: SIZE = Total assets. LTD/TA = Long term debt to total assets ratio. INSDP = %-shareholdings by insiders. INSDPSQ = INSDPsquared. BLOCP = %-shareholdings by blockholders. INSTP = %-shareholdings by institutions. MNC network = NFC or NFR or HERF1 or HERF3. DER/FS = Derivatives’ notional value / Foreign sales. OPT/FS = Options notional value / Foreign sales. FWD/FS = Forwards notional value / Foreign sales. SWAP/FS = Swaps notional value / Foreign Sales. OTH/FS = Other derivatives notional value / Foreign sales. Panel A: The dependent variable, |NBEWR|, is the currency exposure coefficient obtained using the nominal Broad Index and the equally-weighted CRSP market portfolio return. Variable
MNC network (operational hedging ability) is measured by : NFC 1a
Intercept SIZE LTD/TA INSDP INSDPSQ BLOCP INSTP MNC network DER/FS
0.4963 (1.12) -6 2.6×10 (1.07) -1.6458 * (-1.80) 0.0767 (1.49) -0.0014 (-1.39) -0.0112 (-1.32) 0.0251 *** (2.85) -0.0373 ** (-2.37) -0.8186 *** (-3.14)
208 0.2451 3.01
FWD/FS SWAP/FS OTH/FS
Adj.-R
2
F-value
1b 0.5424 (1.21) -6 2.1×10 (0.88) -1.6777 * (-1.81) 0.0759 (1.50) -0.0014 (-1.34) -0.0117 (-1.36) 0.0250 *** (2.83) -0.0389 ** (-2.38)
2.0734 (0.83) -1.2053 *** (-2.72) -0.4394 (-0.66) -0.8145 *** (-3.37) 208
OPT/FS
N
NFR 2a 0.6459 (1.50) -6 1.3×10 (0.57) -1.4955 * (-1.75) 0.0758 (1.51) -0.0014 (-1.40) -0.0107 (-1.30) 0.0255 *** (2.87) -0.1636 ** (-2.51) -0.8120 *** (-3.41)
HERF1 2b
0.6913 (1.58) -6 0.5×10 (0.25) -1.5120 * (-1.76) 0.0755 (1.49) -0.0014 (-1.36) -0.0112 (-1.34) 0.0254 *** (2.85) -0.1683 ** (-2.54)
208
1.7527 (0.73) -1.3112 *** (-2.89) -0.3843 (-0.63) -0.7909 *** (-3.58) 208
0.2479
0.2495
3.19
2.98
3a -0.1112 (-0.21) -6 -0.7×10 (-0.31) -1.9382 ** (-2.18) 0.0748 (1.63) -0.0015 (-1.53) -0.0078 (-1.12) 0.0239 *** (2.85) 1.0320 * (1.75) -0.9758 *** (-3.60)
HERF2 3b
-0.0892 (-0.17) -6 -1.7×10 (-0.84) -1.9604 ** (-2.17) 0.0753 (1.62) -0.0015 (-1.51) -0.0083 (-1.17) 0.0240 *** (2.83) 1.0342 * (1.74)
4a -0.5314 (-0.80) -6 0.2×10 (0.08) -1.5156 * (-1.88) 0.0769 (1.58) -0.0015 (-1.49) -0.0098 (-1.25) 0.0253 *** (2.86) 1.1801 ** (2.11) -0.9842 *** (-3.61)
4b -0.5106 (-0.76) -6 -0.8×10 (-0.37) -1.5326 * (-1.87) 0.0772 (1.57) -0.0015 (-1.47) -0.0102 (-1.29) 0.0253 *** (2.84) 1.1870 ** (2.11)
208
-0.0831 (-0.04) -1.3757 *** (-3.02) -0.0682 (-0.10) -0.9696 *** (-4.13) 208
208
0.4658 (0.21) -1.4386 *** (-3.05) -0.3236 (-0.52) -0.8798 *** (-4.09) 208
0.2521
0.2396
0.2416
0.2496
0.2516
3.19
2.64
2.83
2.76
3.02
30
Table 5 cont’d. Panel B: The dependent variable, |RBEWR|, is the currency exposure coefficient obtained using the real (price-weighted) Broad Index and the equally-weighted CRSP market portfolio return. Variable
MNC network (operational hedging ability) is measured by : NFC
Intercept SIZE LTD/TA INSDP INSDPSQ BLOCP INSTP MNC network DER/FS
1a
1b
2a
2b
3a
0.5318 (1.09) -6 3.1×10 (1.26) -1.6410 ** (-1.99) 0.0781 (1.52) -0.0012 (-1.23) -0.0135 (-1.59) 0.0268 *** (3.03) -0.0445 *** (-2.71)
0.6461 (1.41) -6 2.1×10 (0.91) -1.4335 * (-1.86) 0.0783 (1.58) -0.0013 (-1.29) -0.0123 (-1.53) 0.0274 *** (3.08) -0.1859 ** (-2.87) -0.8403 *** (-3.24)
0.7000 (1.50) -6 1.3×10 (0.59) -1.4607 * (-1.87) 0.0777 (1.56) -0.0012 (-1.25) -0.0129 (-1.57) 0.0273 *** (3.06) -0.1920 *** (-2.91)
-0.2079 (-0.36) -6 -0.2×10 (-0.08) -1.9273 ** (-2.24) 0.0777 * (1.71) -0.0014 (-1.48) -0.0091 (-1.32) 0.0256 *** (3.01) 1.1440 * (1.86) -0.9727 *** (-3.54)
208
2.6743 (1.01) -1.2290 *** (-2.70) -0.4376 (-0.61) -0.7982 *** (-2.89) 208
0.2628 3.34
FWD/FS SWAP/FS OTH/FS
Adj.-R
2
F-value
HERF1
0.4771 (1.00) -6 3.7×10 (1.46) -1.6042 ** (-1.97) 0.0792 (1.56) -0.0013 (-1.28) -0.0129 (-1.55) 0.0270 *** (3.05) -0.0425 *** (-2.70) -0.7903 *** (-2.98)
OPT/FS
N
NFR
208
2.2925 (1.01) -1.3506 *** (-2.97) -0.3740 (-0.58) -0.7723 *** (-3.00) 208
0.2669
0.2679
2.96
3.28
HERF2 3b
-0.1830 (-0.32) -6 -1.3×10 (-0.59) -1.9528 ** (-2.24) 0.0781 * (1.70) -0.0014 (-1.46) -0.0097 (-1.37) 0.0256 *** (2.99) 1.1472 * (1.86)
4a -0.6392 (-0.89) -6 0.7×10 (0.29) -1.4588 * (-1.93) 0.0810 * (1.69) -0.0014 (-1.44) -0.0112 (-1.45) 0.0269 *** (3.00) 1.2454 ** (2.13) -0.9857 *** (-3.57)
4b -0.6163 (-0.85) -6 -0.3×10 (-0.14) -1.4780 * (-1.93) 0.0814 * (1.68) -0.0014 (-1.41) -0.0117 (-1.50) 0.0269 *** (2.97) 1.2541 ** (2.12)
208
0.1781 (0.09) -1.4267 *** (-3.10) -0.0160 (-0.02) -0.9793 *** (-3.92) 208
208
0.7311 (0.36) -1.4950 *** (-3.22) -0.2928 (-0.44) -0.8877 *** (-3.55) 208
0.2717
0.2534
0.2561
0.2600
0.2628
2.94
2.90
2.67
3.05
2.80
31
Table 6 Relationship Between Absolute Exposure and Firmwide Risk Management This table reports the relationship between absolute exchange rate exposure and the combined usage of financial and operating hedges, our test for firmwide risk management. Weighted least squares regressions with robust standard errors (White (1980)) are used to control for heteroskedasticity. The weighted least squares procedure weights by the exposure coefficient’s standard error. Significance levels are indicated as follows: *** 1%, ** 5%, * 10%. The independent variables are: SIZE = Total assets. LTD/TA = Long term debt to total assets ratio. INSDP = %-shareholdings by insiders. INSDPSQ = INSDP-squared. BLOCP = %-shareholdings by blockholders. INSTP = %-shareholdings by institutions. We use two indicator measures of combined hedging, FIRMWIDE1 and FIRMWIDE2. FIRMWIDE1 takes the value of 2 if both NFC and DER/FS are larger than the corresponding sample medians, the value of 1 if one of the two variables’ value is greater than the median, and the value of 0 otherwise. FIRMWIDE2 is similar to FIRMWIDE1 but NFR is used instead of NFC. Dependent variable = |NBEWR|
Dependent variable = |RBEWR|
Variable FIRMWIDE1
1a
FIRMWIDE2
1b 0.1411 (0.24)
1c 0.3141 (0.72)
FIRMWIDE1
1d 0.1414 (0.25)
2a 0.3512 (0.76)
FIRMWIDE2
2b 0.1356 (0.23)
2c
Intercept
0.3747 (0.84)
SIZE
-2.01×10 (-0.11)
LTD/TA
-1.6219 * (-1.80)
-1.8480 ** (-2.10)
-1.4437 * (-1.68)
-1.6716 * (-1.94)
-1.5278 ** (-1.99)
-1.8038 ** (-2.17)
-1.3561 * (-1.79)
INSDP
0.0776 (1.46)
0.0692 (1.53)
0.0791 (1.52)
0.0718 (1.59)
0.0787 (1.52)
0.0701 (1.58)
0.0807 (1.60)
INSDPSQ
-0.0013 (-1.32)
-0.0013 (-1.35)
-0.0013 (-1.35)
-0.0013 (-1.37)
-0.0012 (-1.20)
-0.0011 (-1.24)
-0.0012 (-1.20)
BLOCP
-0.0103 (-1.24)
-0.0092 (-1.28)
-0.0094 (-1.19)
-0.0088 (-1.24)
-0.0120 (-1.47)
-0.0110 (-1.54)
-0.0120 (-1.47)
INSTIT
0.0243 *** (2.81)
0.0248 *** (2.96)
0.0253 *** (2.92)
0.0255 *** (3.03)
0.0263 *** (3.09)
0.0267 *** (3.23)
0.0275 *** (3.23)
FIRMWIDE
-0.5365 *** (-3.17)
-0.3651 * (-1.73)
-0.5666 *** (-3.27)
-0.4219 * (-1.89)
-0.6346 *** (-3.74)
-0.4963 ** (-2.38)
-0.6642 *** (-3.88)
208
208
208
208
208
208
208
0.2437
0.2459
0.2550
0.2631
0.2725
0.2832
0.2858
0.2919
2.85 ***
2.67 ***
3.09 ***
3.02 ***
3.69 ***
3.22 ***
3.72 ***
3.29 ***
-7
2.59×10 (0.14)
-7
-0.01×10 (-0.01)
-7
3.42×10 (0.19)
-7
9.80×10 (0.53)
-7
11.00×10 (0.63)
0.6739 (1.01)
0.5453 (0.78)
0.6562 (0.96)
DER/FS
-0.3974 (-1.32)
-0.3276 (-1.13)
-0.1855 (-0.58)
Adjusted-R F-value
2
2d
0.1246 (0.21) 11.70×10-7 (0.68) -1.5758 * (-1.90) 0.0739 * (1.68) -0.0011 (-1.27) -0.0104 (-1.48) 0.0276 *** (3.35) -0.5545 ** (-2.53) 0.5029 (0.71) -0.1196 (-0.38) 208
HERF1
N
0.2773 (0.62) -7
11.60×10 (0.64)
-7
Appendix 1 Currency Indexes and Trade Weights The table presents the 1997 trade weights for the U.S. dollar indexes used in the study. Index weights are updated annually and previous years index weights are available at the Federal Reserve Board’s web site (http://www.federalreserve.gov). Weights are listed in percent. Components may not sum to totals because of rounding. Country or Region Canada Euro area (see below) Japan Mexico China United Kingdom Taiwan Korea Singapore Hong Kong Malaysia Brazil Switzerland Thailand Australia Indonesia Philippines Russia India Sweden Saudia Arabia Israel Argentina Venezuela Chile Columbia
Broad Index 17.3 16.4 14.6 8.6 6.6 4.6 3.9 3.7 3.1 2.8 2.4 1.9 1.8
Major Index 30.3 28.7 25.6
OITP Index
19.9 15.3 8.0 9.1 8.6 7.2 6.6 5.5 4.4 3.2
1.7 1.5 1.3 1.2 0.9 0.9 0.9 0.9 0.9 0.6 0.6 0.5 0.5
3.9 2.6 3.0 2.7 2.2 2.2 1.6 2.1 2.1 1.5 1.4 1.3 1.1
Total
100.0
100.0
Euro-area countries Germany France Italy Netherlands Belgium/Luxenbourg Spain Ireland Austria Finland Portugal Total Euro-area
5.6 2.9 2.5 1.5 1.4 0.8 0.7 0.4 0.3 0.1 16.4
9.9 5.0 4.5 2.7 2.5 1.4 1.3 0.7 0.6 0.2 28.7
33
100.0
