Exchange rate risk management What can we learn from financial crises(1)

1、#L-H. Lan erat / Economic Modelling45 (2015) 187-192Con tents lists available at ScienceDirectELSEVIEREconomic Modellingjournal homepage: www.elsev/ier.corn/locaie/ecmodCrossMarkExchange rate risk management: What can we learn from financial crises?Li-Huei Lana, Chang-Chih Chen b,*. Shuang-Shii Chua

2、ng3a Department of Business Administration, National Cheng Kung University. Taiwanb International Center of Financial ResearchJiang-Xi Normal University (Yao-Hu Campus). ChinaARTICLE INFOABSTRACTArtide histoiy:Accepted 19 November 2014Available online 18 December 2014Keywords: Exchange rate risk Fin

3、ancial crisis Hedging strategySince most present studies on exchange rate risk have pointed out that it does exist, firms need to hedge all cu rencies in use However, by examining the discrepancy between cost-side and levcnuc-sidc exposures across two major financial crises for Taiwanese firms, we f

4、ind that the exposure mainly comes from the revenue side during the subprime crisis, while that conies from the cost side during the Asian crisis. Our results offer an applicable implication lhal as long as the cost-based or revenue-based hedging strategy can acquire same hedging benefit and effecti

5、vely reduce hedging cost.2014 Elsevier B.V. All rights reserved.http:/dx.doix)rg/10.1016,.j.econmod.2014.11.0180264-9993/ 2014 Elsevier B.V. All rights reserved.1. IntroductionThe 2007-2008 subprime mortgage crisis destroyed the confidence of market investors, leading to a significant collapse in st

6、ock and exchange rate markets worldwide 1 ncvitably. these combined impacts affected firm exposure to exchange rates, as well as its determinants When crises produce sudden shocks. firms formulate or adjust hedging strategies to manage exchange rate risk efficiently. Today, a wealth of literature ex

7、ists, empirically documenting the evidence that financial crises produce structural changes in the exchange rate (see. for instance. Jeon and Seo, 2003; Benson and Faff. 2004： Kaminsky. 2006; Mazouz et al., 2009; Melvin and Taylor. 2009). Given the reality of exchange rate risks, the assumption is t

8、hat firms need to hedge all currencies in use. Nonetheless, adopting any hedging strategy is not without cost. If firms, however, can determine the source of exposure to be revenueside or cost-side, they can effectively manage their exchange rate risk with lower hedging costs.The subprime mortgage c

9、risis followed the Asian financial crisis in 1997. These are the two most significant financial events in thei门 espec tive decades, and they have already received considerable international attention. The two debacles have in common their destructive impact on stocks. However, there arc two importan

10、t ways in which the two crises differ. First, the Asian financial crisis was regional, whereas the impact of the subprime mortgage crisis was global. Second. the cause of the global subprime crisis was the clustering of default on subprime mortgage contracts embedded in asset-backed security derivat

11、ives, whereas the Asian crisis was induced by the so-called double mismatch problem.* Corresponding author at: No. 99. Zi-Yang Road. Nanchang City. Jiang Xi. China. TeUfax： 4-86-mail address: .tw (C.-C. Chen) See Goldstein (1998) for more details about MdouWe mismat

12、ch*1.2 We have also cried the import ratio as the criteria of sample selection but encountered difficulties in finding it. such as data unavailability and trial with other proxy variables. Since present literature on exchange rate risk analyzes only one event, these approaches cannot provide a satis

13、factory answer to the interesting question of how differences between the two crises have affected the dynamic of the returns (in terms of both the exchange rate and a firms stock price). Hcncc. this paper compares the firm*s exchange rate exposure and the determinants of this exposure during the tw

14、o crises.This study contributes to the body of literature concerning the mark agenient of exchange rate risk in practice. A novelty in this study is the introduction of a firm*s exposure as based on its cost and revenue sides. The main objective is to present empirical evidence for the discrepancy b

15、etween cost-side and revenue-side exposures across two major financial crises, and assess risk management with lower hedging costsInspired by the asymmetry between the two crises, we divided this study into three time periods: (i) the period of the Asian financial crisis (July 1997 to December 1998)

16、; (ii) the period of the subprime mortgage crisis (July 2007 to December 2008); and (iii) the intervening non-crisis period (January 1999 to June 2007). This study is aimed at answering the following questions: (i) Are there any significant discrepancies in firms* exchange rate risks between the sub

17、prime (global) crisis and the Asian (regional) crisis? (ii) How do the different geographical scopes of the two crises explain difFercnces in the exposure to exchange rate risk, as assessed by a cost-/revenue-based exposue analysis? (iii) What is the interaction between the type of industry and the

18、char acteristics of exchange rate risk of a firm representing that industry, controlling for the coverage of the crisis?2. Literature reviewExchange rate risk is generally defined in literature as the sensitivity of a firm*s value to fluctuations in exchange rates. To understand this effect theoreti

19、cally, researchers typically model the structure of the joint dynamic effects of profits and exchange rates with reference to a specific setting defined by market supply and demand (see, for instance, Bodnar and Marston. 2002： Bodnar et al.t 2002; Bartram et al. 2010).Then, the value of exchange rat

20、e risk is derived using a differentiation operator. Most exchange rate risk theories operate on the plausible assumption that firms w让h foreign operations face the greatest exposure to exchange rate riskThe issues surrounding exchange rate risk have inspired a number of empirical studies In these st

21、udies, the estimates of exchange rate risk are primarily obtained by a time-series regression w让h the return on the per-share stock price as the dependent variable, changes in the ex change rate as the independent variable, and the return on the market index as the control variable. An early example

22、 of how to derive an empirical estimate of exchange rate risk is given byjorion (1990). whose analysis was based on U.S. data. Jorion (1990) discovered that the ratio of multinationals w让h significant exposure to the total sample was fairly low (ca. 10-25%) and positively correlated with a firms deg

23、ree of foreign involvement. This low ratio has been found to hold for other single countries and regions, including Japan (He and Ng, 1998), Taiwan (Chiao and Hung, 2000). Germany (Bartram. 2004), the AmericanEuo zone (Bartram and Karolyi, 2006). and the American- European-Asian zone (Doidge et aL.

24、2006).The results of the ratio test, of significant exchange rate risk, reported in most empirical studies are inconsistent with the traditional theories, maintaining that firms with foreign operations must experience considerable exchange rate risk. This is the so-called exchange rate exposure puzz

25、le (Bartram and Bodnar. 2007). and several attempts to solve it are present in liteiature. These solutions involve the following hypothesized effects: (i) the mixture offset effect, a tradeoff in using the average value of exposure as the measure of exposure level, while not distinguishing between p

26、ositive and negative exposure (see for instance, He and Ng, 1998; El-Masry and Abdel-Salam, 2007)： (ii) the industrial grouping effect, the underestimation of the exposure ratio across industries caused by the inability of the analysis to cope with het crogcncity in industries* foreign in volvcmcnt

27、(Miller a nd Rcucr, 1998). To date, as evidence indicates (hat the abovementioned hyix)theses are commonly weak and not supported by robust tests, the exposure puzzle remains unsolvedAchieving a deeper understanding of a firms exposure to currency fluctuations requires the exploring of exchange rate

28、 risk determinants. This study summarizes the arguments in literature (including Jorion, 1990; Nance et al. 1993： He and Ng. 199& Chiao and Hung, 2000) to present five exposure factors: (i) firm size (as a proxy for economies of scale in hedging costs); (ii) the ratio of foreign sales to total sales

29、 (as a proxy for the degree of foreign invoIvement); (iii) quick ratio (as a proxy for liquidity); (iv) book to market ratio (as a proxy for growth opportunities); and (v) debt ratio (a proxy for the likelihoods of financial distress). However, only consistent evidence on the positive correlation be

30、tween the exposure level and the extent to which a firm is involved in foreign operations is revealed.Whereas, the earlier studies primarily focus on individual exposure behavior, recent studies have shifted attention to the external effects ir) duced by major financial events. For example, Allayann

31、is et al. (2001) analyzed the exchange rate exposure of firms in eight East-Asian tries during the Asian financial crisis. Tliey found that the onset of the nancial crisis produced structural changes in the currency fluctuations and the pattern of exposure Likewise. Parsley and Popper (2006) found s

32、ignificant increments in exchange rate risk during the Asian cri- sis. using data from the Asia-Pacific zone Another financial event often addressed in academic papers on foreign exchange rate is the reform of the exchange rate mechanism Motivated by the breakdown of the famous Bretton Woods system,

33、 Bartov et al. (1996) discovered apparent sharp increases in U.S. firms* exchange rate risk after the announcement of the floating exchange rate. El-Masry and Abdel-Salam (2007) found that adoption of the European exchange rate increased U.K. firms* exposure to exchange riskDespite the substantial d

34、evelopment of exchange rate risk, most previous studies have analyzed only one event. Thus, whether there are any significant differences in the effects of two crises on exposure-related behavior is still unknown. How differences in the coverage of crises explain the asymmetry in the impact of the c

35、rises on such behavior is another interesting question that helps motivate the present study.3. Research data and methods3.1 Sample selection and data collectionFollowing Jorion (1990) and He and Ng (1998), we selected Taiwanese multinational corporations with an export ratio of at least ten percent

36、 in the full sample period from 1997 to 2008? As a result. 204 firms are included in the sample, two thirds of which (134) belong to eight major industries: 32 in fibers and textiles. 17 in computer equipment, 16 in semiconductors. 16 in plastics, 15 in iron and steel, 15 in electronic components.12

37、 in chemicals, and 11 in electrical equip ment. Relevant stock prices and financial data are obtained from the Taiwan Economic Journal (TEJ) database A summary of descriptive statistics of the sample data is compiled in Table 1. The firms with the largest average size ($5975 million), quick ratio (2

38、.21), and book- to- market ratio (67%) all belong to the semiconductor industry： the iron- and-steel firms had the highest average long-term debt-to-equity ratio (46%); the computer equipment firms had the highest average export ratio (86%).Data on returns from the exchange rate were retrieved from

39、direct quotations on the Oanda website (/). To con struct the trade-weighted exchange rate index, we took the weighted average of the five major bilateral exchange rates, defined as the number of New Taiwan Dollars per unit of the foreign currencies of China, Hong Kong. Japan, Kor

40、ea, and the U.S., respective!y. These weights. which represent each countrys proportion of the five countries* total trade with Taiwan (see Table 2 for details), were computed using trade data from the Taiwan Directorate General of Customs. The data were ranked according to (he following three crite

41、ria: imports, exports. and total trade. This procedure provides a thorough understand ing of the discrepancies between a firms cost-side and revenue-side exchange rate risk caused by the asymmetry in the geographical scope of the crisis. To further investigate whether there were structural chang es

42、in a firnYs exchange rate risk during the crisis period. as well as the differential impact of the two crises on stock prices and the exchange rate markets, the analysis period was divided into three subperiods: July 1997 to December 1998, January 1999 to June 2007, and July 2007 to December 200&32.

43、 Exchange race riskDumas (1978). Hodder (1982), and Adler and Dumas (1984) define exchange rate risk as the effect of exchange rate changes on a firms value. Because a firm usually undertakes hedging activities for expected currency fluctuations the analysis requires that first the unexpected change

44、s are separated from total changes in (he exchangeale. Following El-Masry and Abdel-Salam (2007). this can be achieved via a timeseries autoregression:ER( = 60 + 5|ERf_ + UERt(1)where ER( is the rate of total return on exchange rate at time t and the residual term UER( measures the unexpected change

45、s in the exchange rate .In estimating the exposure to exchange rate risk for firms this study borrows the two-factor model from Jorion (1990):+ 畑MRf + 印，r = 1 广J(2)L-H. Lan et aL / Economic Modelling 45 (2015) 187-192191IndustrylogS/ZEQRDEBMEXPRMeanS.D.MeanS.D.MeanS.D.MeanS.D.MeanS.D.Plastic (16/7.8

46、%)510988551.470.820.380.110.620.110.420.21Fib. & text. (32/15.5%)111819631.161.290.380.120.620.120.430.24Elec. mach. (11/5.4%)161329801.080.850.440.120.560.120.550.27Chemical (12/5.9%)7827181.842.920.360.110.640.110.430.17Iron& steels (15/7.4%)258756370.650.450.460.160.540.160.360.23Semi. (16/7.8%)5

47、975106602.211.620330.130.670.130.550.18Com. equip. (17/8.3%)336636071.410.620.430.090.570.090.860.18Elec. comp. (15/7.4%)116315781.390.550.380.110.620.110.700.21Other (70/34.5%)182330701.421.190.380.130.620.130.490.31Total (204/100%)261543411.401.150.400.120.610.120.530.22Table 1Descriptive statisti

48、cs of sample.Note: In parentheses, the left-side is the number of firms and the right-side is its percentage. logS/ZE is the log-size of the firm, QR is the quick ratio, DE is the debt-to-equity ratio, BM is the book-to-market ratio for the equity, and EXPR is the export ratio.where SRit is the retu

49、rn rate of firm is stock, MRt is the return rate of stock market index, and 卩“ is firm is exposure value.We are now prepared to introduce the model for testing the determinants of foreign currency exposure. Based on prior studies (such as Jorion, 1990; Nance et al., 1993; He and Ng, 199& Chiao and H

50、ung, 2000), we used our cross-sectional regression to generate five exposure factors:伤：=a。+ a 1 log SIZE j + cIqEXPR： + cQR： + q5DE+ 巧(3)where log SIZE： is the log-size of the firm, EXPR： is the export ratio, Q& is the quick ratio, BM： is the book-to-market ratio for the equity, and DE： is the debt-

51、to-equity ratio. These five factors in the analysis are respectively treated as the proxy for economies of scale in hedging costs, the degree of foreign involvement, firm liquidity, growth opportunities, and the likelihood of financial distress3.3. RobustnessThis research conducts the robustness tes

52、ts in two ways. First, to examine whether the right-hand-side variables in Eq. (3) display symmetric effects on both the positive and negative By (to control Umixture offset0 effect), we add to Eq. (3) the dummy variable D, following He and Ng (1998) and El-Masry and Abdel-Salam (2007), which is ass

53、igned the value 1 if is positive and 0 otherwise:B、i = b0D + D logS/ZE ： + b2DEXPRi + b3DQRI + b4DBMf + b5DDEf + -0(1D) + bdi(lD)logSJZEj + bd2(lD)EXPR： + 财(1D)QRj + bd4(lD)BMj + bd5(lD)DEj + 3.(4)Second, as suggested by Miller and Reuer (1998), to rule out possible bias resulting from grouping the

54、industries, we adopted an approach similar to that used for Eq. (4) by rewriting Eq. (3):Eli = 51D + Q2D2 + Q3D3 +Q4D4 + cd5D5 + cd6D6 + cd7D7 + cfl8D8 + c ogSIZEi + c2EXPRj + c3(2Rj + c4BMj + c5DEj + 比(5)Table 2Top-five Taiwanese trade partner countries.U.S.JapanHong KongChinaKoreaTotalImport ($ mi

55、llions)257,418434,26024,035160,777116,981993,471Percentage25.91%43.71%2.42%16.18%11.78%100%Export ($ millions)357,038163,907377,397301,68356,8611,256,886Percentage28.41%13.04%30.03%24.00%4.52%100%Total ($ millions)614,456598,167401,431462,461173,8422,250,358Percentage27.30%26.58%17.84%20.55%7.73%100

56、%Note: Import volume plus export volume equals the total trade volume.where the control variable D： = 1.8 is assigned the value 1 depending on the industry (i) as follows: for plastics, D2 for fibers and textiles, D3 for electrical equipment, D4 for chemicals, D5 for iron and steel, D6 for semicondu

57、ctors, D7 for computer equipment, and D8 for electronic components. When all the dummy variables equal 0, D： = 1.8 is labeled “other”.4. Empirical resultsThe statistical analyses were conducted in three stages In the first stage, the firms* total exchange rate risks during the two crisis periods wer

58、e compared. In the second stage, to further understand the consequences of the asymmetrical impact of the two crises on exposure behavior, we incorporated the distinction between import-based and export-based data in constructing the trade-weighted exchange rate index and performing the exposure analyses. In the third stage, we iden tified