Exchange rate regimes and foreign exchange exposure: The case of emerging market firms

Emerging Markets Review 21 (2014) 156–182

Contents lists available at ScienceDirect

Emerging Markets Review journal homepage: www.elsevier.com/locate/emr

Exchange rate regimes and foreign exchange exposure: The case of emerging market firms Min Ye a,⁎, Elaine Hutson b, Cal Muckley c a b c

College of Finance and Statistics, Hunan University, China Department of Banking and Finance, Monash University, Australia UCD School of Business and Geary Institute, University College Dublin, Ireland

a r t i c l e

i n f o

Article history: Received 23 May 2014 Received in revised form 24 July 2014 Accepted 2 September 2014 Available online 10 September 2014 JEL classification: E42 F31 F33 Keywords: Foreign exchange exposure Exchange rate regimes Emerging markets

a b s t r a c t We investigate the influence of exchange rate regimes on the foreign exchange exposure of emerging market firms. Using a sample of 1523 firms from 20 countries for the period December 1999 to December 2010, we find that about half of the firms are significantly exposed to exchange rate fluctuations. We find that non-floating exchange rate arrangements are associated with more widespread exposure as well as a greater magnitude of firms' exposure. Cross-sectional analyses suggest that the exchange rate regime is an important determinant of firm-level exchange rate exposure for emerging market firms, and that pegged exchange rate regimes amplify exposure. This result holds after controlling for a wide range of potential determinants of firm-level and country-level foreign exchange exposure. Our findings suggest that exchange rate regime matters at the micro as well as the macro level; non-floating regimes fail to protect firms from exchange rate exposure. © 2014 Elsevier B.V. All rights reserved.

1. Introduction According to Calvo and Reinhart (2002) there is an epidemic of ‘fear of floating’ among emerging economies. This ‘epidemic’ appears to have been exacerbated by the financial crisis; emerging economies have recently shifted toward more stable exchange rate regimes and away from

floating arrangements (IMF, 2011). Proponents of ‘fear of floating’ often focus on the ability of stable exchange rates to facilitate bilateral trade (Frankel and Rose, 2005; Klein and Shambaugh, 2006), reduce the output cost associated with exchange rate fluctuations (Lahiri and Végh, 2002), and reduce inflation (Ghosh et al., 2002). Yet

⁎ Corresponding author. E-mail addresses: [email protected] (M. Ye), [email protected] (E. Hutson), [email protected] (C. Muckley)

http://dx.doi.org/10.1016/j.ememar.2014.09.001 1566-0141 © 2014 Elsevier B.V. All rights reserved.

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

157

opponents emphasize that pegged exchange rate regimes are associated with weaker economic growth (Cruz-Rodriguez, 2013; Levy-Yeyati and Sturzenegger, 2003; Petreski, 2009) and greater financial fragility (Burnside et al., 2001; Chang and Velasco, 2000; Eichengreen and Hausmann, 1999). This ‘fear of floating’ debate focuses on the effects of the exchange rate regime on macroeconomic outcomes. In this paper, we provide insight into the debate from a microperspective — by exploring the effects of exchange rate regimes on firm-level foreign exchange exposure. There are three main reasons why the exchange rate regime might affect firms' foreign exchange exposure. First, a country's exchange rate arrangement is an important determinant of exchange rate volatility — fixed regimes are associated with lower exchange rate volatility (Rossi, 2009). Volatility engenders firms' direct foreign exchange exposure — which arises for firms with foreign assets and liabilities as well as expected future foreign currency cash flows. Volatility also affects firms' exposure indirectly. Several scholars have argued that exchange rate volatility increases the cost of hedging (Arteta, 2005; Eichengreen and Hausmann, 1999; McKinnon, 2000), which may result in less hedging activity and thus higher foreign exchange exposure. Second, the ‘moral hazard hypothesis’ advanced by Eichengreen and Hausmann (1999) suggests that pegged exchange rates can be viewed as an implicit government guarantee — which may diminish firms' incentives to hedge and promote the use of unhedged foreign currency debt (Burnside et al., 2001; Fischer, 2001; Schneider and Tornell, 2004). Third, under fixed or heavily managed exchange rate regimes, firms (or their creditors) may regard an extreme devaluation or appreciation as a rare event, and therefore underestimate, neglect or even are not aware of the associated exchange rate risk (Kamil, 2006, 2012). They are thus less likely to actively engage in hedging to mitigate their foreign exchange exposure. It is likely, therefore, that a country's exchange rate regime has a major impact on the foreign exchange exposure of its firms. However, there have been very few studies of this issue. Parsley and Popper (2006) examined firms' foreign exchange exposure in several Asian markets for the period January 1990 to March 2002, and found that during dollar peg periods firms were highly exposed to movements in the dollar, and in some countries more firms were significantly exposed to the dollar with a peg than without one. Patnaik and Shah (2010) studied the foreign exchange exposure of the 100 most liquid Indian stocks from April 1993 to March 2008, and found that firms experienced higher exposure in periods when the currency was less flexible. While these studies provide some clues as to the relation between exchange rate regimes and firms' foreign exchange exposure, they did not examine how the country's foreign exchange regime affected firms' exposure, and nor did they look at the effect of exchange rate regimes on the extent to which firms are exposed. Given the recent shift toward less flexible exchange rate systems in emerging markets, it is timely to revisit the issue of exchange rate arrangements and firm-level foreign exchange exposure. This paper has three main novel features. First, we are the first to closely study the relation between exchange rate regimes and firm-level foreign exchange exposure. That is, we examine firms' exchange rate exposure across countries with different exchange rate regimes, and we also investigate whether exposure alters when exchange rate arrangements change. Second, we examine whether and by how much exchange rate regimes determine firms' foreign exchange exposure. According to Chue and Cook (2008), Choi and Jiang (2009), Hutson and Stevenson (2010), Aggarwal and Harper (2010) among others, the magnitude of a firm's foreign exchange exposure may be associated with certain firm-specific traits (such as size, growth opportunities, and expected financial distress), and country-specific characteristics (such as trade openness and the corporate governance environment). Controlling for a wide variety of factors that are known to affect exchange rate exposure, we examine whether the exchange rate regime is a significant determinant of emerging market firms' exposure. Third, we use a unique data set that is larger than any used in prior studies. Our data set comprises 1523 firms in 20 emerging markets for the period December 1999 to December 2010. During this period, according to the IMF's database of exchange rate arrangements, 2 of these 20 countries had exclusively pegged exchange rate arrangements, 11 had exclusively floating exchange rates, and 7 alternated between floating and pegged regimes. This diversity of exchange rate systems enables us to examine whether firms' exchange rate exposure differs under floating versus non-floating regimes, as well as how exposure varies with changes in the exchange rate regime in the country in which the firm is based. Further, a series of sub-period analyses allows us to examine how emerging market firms' exposure has changed over time, including during the recent global financial crisis.

158

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

The empirical results yield three key findings. First, we find that the magnitude of exposure is greater for firms in countries with pegged exchange rates than for those with floating regimes, which is consistent with the findings of Parsley and Popper (2006). Second, our cross-market tests of firms in the seven countries that experienced both floating and non-floating regimes during the sample period show that a switch to non-floating exchange rate arrangements is associated with an increase in firm-level foreign exchange exposure — both a higher proportion of significantly exposed firms and greater magnitude of exposure. Third, our cross-sectional analysis suggests that the exchange rate regime is a statistically significant determinant of firm-level exchange rate exposure, and that pegged exchange rate regimes are consistently associated with greater exposure for emerging market firms. This result holds after controlling for a wide range of potential determinants of firm-level and country-level foreign exchange exposure. Our findings have implications for the ongoing debate about the merits and disadvantages of floating versus non-floating exchange rates. Exchange rate regime matters at the micro as well as the macro level. Non-floating regimes fail to protect firms from exchange rate exposure — and in fact can exacerbate exposure. This may be explained by the fact that pegged exchange rates could lull firms into taking on too much unhedged foreign currency debt that leaves them exposed to external shocks (the ‘moral hazard hypothesis’). It is also possible that under non-floating exchange rate regimes, firms (or their creditors) may be less aware of exchange rate risk, and may therefore be less likely to actively engage in hedging to mitigate their foreign exchange exposure (Kamil, 2006). Our findings also have a natural implication for firms' management and regulators, that the adoption of tightly managed exchange rate regimes would induce higher firm-level exchange rate exposure, and thereby require a more cautious management of the exposure in the corporate/private sector. The remainder of our paper is organized as follows. Section 2 reviews the relevant literature. Section 3 estimates the exchange rate exposure of the full sample of firms. Section 4 examines the impact of the exchange rate regime changes on firms' exchange rate exposure, by studying a sub-sample of firms in countries which experienced both floating and non-floating exchange rate regimes. Section 5 examines, using a pooled regression analysis, whether and how the extent of the foreign exchange exposure is related to the country's exchange rate arrangements. Section 6 summarizes and concludes. 2. Foreign exchange exposure and exchange rate regimes: theory and evidence The foreign exchange exposure of a firm is defined as the sensitivity of its economic value to exchange rate changes (Hekman, 1983). A plethora of studies in the international business and finance literatures have explored firm-level foreign exchange exposure, with most focusing on firms in developed countries with floating exchange rate regimes. In contrast, research on the exchange exposure of firms in emerging markets, and particularly in countries with non-flexible exchange rate regimes, is sparse. 2.1. Firm-level foreign exchange exposure in developed and emerging markets Studies examining firm-level exchange exposure in developed countries have met with limited success in documenting significant exposure; this is the so-called ‘foreign exchange exposure puzzle’ (see Bartram and Bodnar, 2007, for a review). Focusing on the US multinational corporations (MNCs), Jorion (1990), Amihud (1994), Bartov and Bodnar (1994) and Shin and Soenen (1999) found little or no relation between changes in exchange rates and the value of the firm. Many non-US studies have similar findings. Loudon (1993) found that 11% of Australian firms were significantly exposed to the foreign currency value of the Australian dollar. Nydahl (1999) found that 17% of Swedish firms were significantly exposed, and using a sample of German firms Bartram (2004) found that 7.5% of the 373 non-financial firms in his sample were significantly exposed. He and Ng (1998) found that a more substantial 26% of Japanese MNCs had significant exchange rate exposure. Large cross-country studies yield similar findings. For 12,821 non-financial firms in 20 countries, Bartram and Karolyi (2006) found few firms with significant exposure, and Doidge et al. (2006) found that only 8% of firms 17,929 firms from Europe, Asia and North America were significantly exposed. Using a sample of 3788 firms from 23 developed countries, Hutson and Stevenson (2010) found significant exposure for 11% of firms. Of the few studies that have been conducted on firms in emerging markets, most have found much higher rates of significant exposure than in studies of developed market firms. Dominguez and Tesar

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

159

(2001)'s multi-country study of firm-level exchange rate exposure included two emerging markets — Chile and Thailand. They found that 19% of Chilean firms and 20% of Thai firms were significantly exposed over the period 1980 to 1999. Kiymaz (2003) found that nearly half (51 of 109) of his sample of Turkish firm experienced statistically significant exposure to exchange rate movements. Parsley and Popper (2006) found that over the period from January 1990 to March 2002 more than 40% of firms from Indonesia, Korea, Malaysia and the Philippines, and 19% of Thai firms, were found to be significantly exposed. In the largest multi-country study on emerging market foreign exchange exposure so far, Chue and Cook (2008) examined the foreign exchange exposure of 931 firms in 15 emerging markets. They found that on average 11% of firms had significant exposure during the period January 1999 through June 2002, falling to 9% in the period July 2002 to June 2006. Aysun and Guldi (2011) found that a considerable proportion of firms from five emerging markets – 17% in Brazil, 29% in Chile, 53% in Korea, 19% in Mexico and 44% in Turkey – were exposed during the period 1995 to 2006. For six Asian emerging markets, Lin (2011) found that the proportion of firms with significant foreign exchange exposure was higher during the 1997 Asian crisis (10%) and the 2008 global crisis (18%), compared to 9% during the relatively tranquil inter-crisis period. 2.2. Financial fragility and exchange rate regimes Following the financial crises of the 1990s (Mexico in 1994, Southeast Asia in 1997, Russia in 1998, Brazil in 1999, Turkey and Argentina in 2001), a number of studies examined the link between the exchange rate regime and financial stability. As the currency mismatches of banks and firms are seen as a source of financial fragility (Arteta, 2005), mismatches are an important element of the debate about the merits and disadvantages of fixed versus flexible exchange rates. Particularly when there is high liability dollarization (Calvo and Reinhart, 2002), large depreciations can weaken bank and corporate balance sheets and can threaten the stability of the financial system. One view is that the exchange rate volatility associated with floating systems increases the cost of hedging, which may result in less hedging and greater currency mismatches (e.g. Arteta, 2005). Examining deposit and credit dollarization in developing and transition countries, Arteta (2005) finds that floating regimes aggravate currency mismatches in domestic financial intermediation. However, most empirical studies support the alternative view – the ‘moral hazard hypothesis’ (Eichengreen and Hausmann, 1999) – that pegged exchange rates increase currency mismatches because of an implicit exchange rate guarantee (Burnside et al., 2001; Chang and Velasco, 2000; Eichengreen and Hausmann, 1999; Schneider and Tornell, 2004). This implicit guarantee promotes risk-taking — encouraging unhedged foreign currency borrowing and the skewing of financial flows toward the short end. In pegged exchange rate regimes the corporate sector will accumulate of unhedged foreign currency debt that leaves them exposed to a sudden reversion of economic conditions (Fischer, 2001; Mishkin, 1997). Martínez and Werner (2002) analyzed the effect of Mexico's 1994 exchange rate regime change on the currency composition of Mexican firms' corporate debt. They found that moving to a more flexible regime was associated with a reduction in balance sheet currency mismatches. Kamil (2006) extended the sample to include Argentina, Brazil, Chile, Colombia, Mexico, Peru and Uruguay, and found that under floating exchange rate regimes, firms (or their creditors) are more aware of exchange rate risk, and therefore act to mitigate foreign exchange exposure by hedging their foreign currency positions. Similar results were found by Kamil (2012), who took a further look at this issue in six Latin American countries and confirmed that the choice of exchange rate regimes affects firms' foreign currency borrowing behaviors and associated balance-sheet currency mismatches. Using a sample of Brazilian firms, Rossi (2011) found that Brazil's switch to a floating regime improved the match between the currency composition of firms' assets and liabilities. There is one paper to our knowledge argued that the choice of exchange rate regime does not matter as to currency mismatching. Baek (2013) examined the determinants of aggregate currency mismatches using the data set covering 97 countries over 1990–2004, and found that exchange rate regime is a statistically insignificant determinant of currency mismatches. 2.3. Foreign exchange exposure and exchange rate regimes The literature on exchange rate regimes and financial fragility in emerging markets focuses almost exclusively on balance sheet currency mismatches, and tends to ignore cash flow mismatches. Firms

160

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

experience direct foreign exchange exposure when they have contractual and expected future foreign currency cash flows. Firms may also bear indirect foreign exchange exposure — that arises from the competitive environment in which the firm operates (Flood and Lessard, 1986; Marston, 2001). For example, a firm that manufactures and sells locally – that is, it has no international operations or transactions – will be exposed to a strengthening domestic currency as competing imports become relatively cheap. Hutson and Stevenson (2010) argued that their finding of a strong relation between country-specific economic openness and firm-level foreign exchange exposure is due to greater indirect exposure of firms in open economies relative to more closed economies. The limited literature on the exposure of firms in countries with pegged exchange rate regimes shows that a peg can fail to protect firms from sensitivity to currency movements. Using a data set of 531 firms in Indonesia, Korea, Malaysia, the Philippines, Taiwan and Thailand, Parsley and Popper (2006) explored firms' foreign exposure under different exchange rate regimes – dollar peg and non-peg, for the period January 1990 to March 2002. They found that during dollar peg periods, firms were highly exposed to the dollar as well as the other currencies. They also found that more firms were significantly exposed to the dollar with a peg than without one in Malaysia, the Philippines, and Thailand. Parsley and Popper (2006) examined only the proportion of firms significantly exposed; they did not look at the effect of exchange rate regimes on the extent to which firms are exposed. Patnaik and Shah (2010) studied the foreign exchange exposure of the 100 most liquid stocks on Indian stock exchanges from April 1993 to March 2008 — when a dollar peg was in place. By dividing the full period into sub-periods based on the variation in exchange rate volatility, they found that firms experience higher exposure in periods during which the currency was less flexible. This study did not, however, examine the relation between exchange rate arrangements and firms' foreign exchange exposure, because there were no substantial changes in exchange rate regime during the period. 3. Measuring foreign exchange exposure 3.1. Model specifications There are two main approaches to estimate foreign exchange exposure: the single-factor and the two-factor model. The single-factor model regresses the contemporaneous exchange rate change on a stock's return; it yields an estimate of total exposure: X

Rn;t ¼ α n þ βn X t þ εn;t

ð1Þ

where Rn,t denotes the return of firm n in period t; and Xt denotes the exchange rate change. The coefficient βXn is firm n's sensitivity exchange rate movements. The two-factor model of Jorion (1990) is an augmented market model, whereby a stock market return is added to Eq. (1) to control for macroeconomic effects, because exchange rates and stock prices are often affected by the same shocks: m

X

Rn;t ¼ α n þ βn Rm;t þ βn X t þ ε n;t

ð2Þ

where Rm,t is the rate of return of the market portfolio in period t. In this equation, βXn measures firm n's sensitivity to exchange rate movements, independent of the effect that changes in the exchange rate have on the overall market. This approach yields an estimate of the firm's so-called residual exposure (Bodnar and Wong, 2000). The standard implementation of the two-factor model involves the use of a country-specific stock market index. This approach, however, is not ideal in multi-country studies. A measure of total exposure is preferred as it encompasses country-level, macroeconomic determinants of exchange rate exposure as well as firm-level. However, the approach to estimate total exposure (Eq. (1)) overestimates exposure since there are macroeconomic variables that co-vary with exchange rate movements and stock returns (see Muller and Verschoor, 2006a for a review of these issues). Parsley and Popper (2006) and Chue and Cook (2008) include a world market return, Rw,t, to control for worldwide macroeconomic factors: W

X

Rn;t ¼ α n þ βn Rw;t þ βn X t þ εn;t :

ð3Þ

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

161

This approach also has its weaknesses. Research on stock market integration has found that emerging markets exhibit different degrees of integration (or segmentation) with world markets. Bekaert (1995) found that in most emerging markets, global factors account for only a small fraction of the time variation of expected returns in most emerging markets. This finding was confirmed by Korajczyk (1996), who showed that emerging markets are less integrated than developed markets. Although in general the trend is toward greater integration in financial markets (Carrieri et al., 2007; Korajczyk, 1996), the process is surely gradual (Bekaert and Harvey, 2002) and it may reverse at times (Carrieri et al., 2007). For this reason, world market returns may not be a competent proxy for all of the macroeconomic shocks to emerging market firms over our sample period. We run a pre-test by estimating Eq. (3), and find that more than half (55%) of emerging market firms have no significant sensitivity to world market returns.1 We therefore incorporate an index of emerging market returns, Re,t, to ‘soak up’ the effects of any remaining macroeconomic variables that affect emerging market firms: W

E

X

Rn;t ¼ α n þ βn Rw;t þ βn Re;t þ βn X t þ εn;t :

ð4Þ

Several studies (for example, Baillie and Bollerslev, 1989; Bollerslev et al., 1992; Hsieh, 1989; Tse, 1998, among others) have documented conditional heteroskedasticity in asset returns, which results in inefficient parameter estimates as well as biased test statistics in the ordinary least squares regression. We use the Lagrange multiplier test proposed by Engle (1982) to check whether the residuals, εn,t, exhibit time-varying heteroskedasticity, and find that in about 90% of cases, the error terms in Eq. (4), εn,t, are heteroskedastic. We therefore incorporate a conditional variance into the system by adding a GARCH (p, q) process. The regression model becomes: W

E

X

Rn;t ¼ α n þ βn Rw;t þ βn Re;t þ βn X t þ εn;t qffiffiffiffiffiffiffiffiffi With ε n;t ¼ μ n;t σ 2n;t Xp Xq 2 2 2 and σ n;t ¼ γ þ τσ þ φε i¼1 i n;t−i j¼1 j n;t− j

ð5Þ

where σ2n,t denotes the conditional variance of the residuals εn,t; γ, τi and φj are unknown parameters2; and μn,t represents the white noise error term. The Akaike (1973) Information Criterion (AIC) and the Schwarz (1978) Information Criterion (SIC or BIC) are used to determine the optimal GARCH (p, q) model for each firm. We find that both of these criteria select GARCH (1, 1) as the optimal model for almost all of the firms. This is consistent with many empirical studies such as Bollerslev et al. (1992) and Muller and Verschoor (2006b), which show that the GARCH (1, 1) specification is optimal for modeling the variance-generating process of financial time series. We therefore use the GARCH (1, 1) specification for our time-series estimations. 3.2. Data and variable description Table 1 presents the summary information about our sample countries and firms. Our data set includes firms from 20 emerging markets3 over the period January 1999 to December 2010. Columns 1 and 2 of Table 1 present two proxies for country size: GDP and total stock market capitalization, drawn from Datastream at 31 December 2010. We use the IMF exchange rate classifications published in the IMF Annual Report on Exchange Rate Arrangements and Exchange Restrictions (AREAER) (2011). The IMF classifies regimes as [a] hard pegs (exchange arrangements with no separate legal tender, and currency

1

The detailed pre-test results are not reported in this paper. They are available from authors upon request. The unknown parameters are estimated by maximum-likelihood and generated using the Bemdt et al. (1974) algorithm. 3 Following Aggarwal et al. (1999), we use the International Finance Corporation (IFC) classification of emerging markets. As of May 2012, there are 21 emerging markets classified by IFC. We include all of these markets except Taiwan, as information on exchange rate arrangements is not available for Taiwan in IMF reports. 2

162

Table 1 Summary information: countries and firms. Country-level data Country

GDP (US$ bn)

(1)

Market cap. (US$ bn)

(2)

(3)

Exchange rate arrangements

No. of firms (total)

Floating

Peg

(4)

(5)

(6)

12/31/1999–12/31/2010 12/31/1999–12/31/2010

297 (352) 28 (41)

12/31/1999–01/28/2003 02/01/2005–07/31/ 2008 03/12/2009–12/31/2010 12/31/1999–02/25/2008 06/21/2010–12/31/2010 12/31/1999–12/31/2005 01/01/2009–12/31/2010 09/16/2009–12/31/2010 03/01/2006–12/31/2010 12/31/1999–02/21/2001

0.03/0.63 0.00/0.25

Countries with non-float and floating regimes Egypt 218.89 82.49

−0.22/1.11

01/29/2003–01/31/2005 08/01/2008–03/11/2009

Hungary Indonesia Malaysia

128.63 706.56 237.80

27.71 360.39 410.53

−0.01/1.09 −0.13/1.52 0.02/0.55

02/26/2008–12/31/2010 12/31/1999–06/20/2010 01/01/2006–12/31/2008

Peru Russia Turkey

157.05 1479.82 734.36

99.83 1004.52 306.66

0.07/0.65 −0.09/0.94 −0.40/2.33

12/31/1999–09/15/2009 12/31/1999–02/28/2006 02/22/2001–12/31/2010

Countries with Brazil Chile Colombia Czech India Korea Mexico Philippines Poland South Africa Thailand

floating regimes 2087.89 212.74 288.89 192.03 1727.11 1014.48 1035.87 199.59 469.44 363.91 318.52

1545.57 341.58 208.50 43.06 1615.86 1089.22 454.35 157.32 190.23 1012.54 277.73

0.04/2.14 0.03/1.30 0.04/1.44 0.14/0.79 −0.05/0.69 −0.04/1.32 −0.11/1.06 −0.07/0.84 0.04/1.21 −0.07/1.95 0.03/0.78

12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010 12/31/1999–12/31/2010

Market cap. med. (US$ m)

Zero-return ratio (%) Min.

Max.

Med.

(8)

(9)

(10)

460.45 98.28

0.00 13.10

36.48 71.32

13.80 26.07

11 (17)

1549.13

6.55

65.51

14.56

24 (37) 85 (286) 17 (22)

66.97 120.80 3299.51

6.90 12.11 19.55

99.30 99.75 34.39

29.47 57.87 27.74

9 (33) 40 (44) 58 (76)

722.07 4010.75 294.45

22.01 4.09 16.27

93.31 92.26 55.64

55.16 13.58 21.78

53 (54) 26 (30) 15 (27) 7 (10) 152 (156) 307 (340) 28 (32) 11 (21) 226 (280) 94 (112) 35 (37)

2940.57 1992.91 1328.26 1255.53 1647.54 77.71 2220.86 1506.28 68.20 1073.01 1065.05

6.35 5.65 11.90 4.66 3.42 1.47 4.24 16.86 0.00 5.28 14.89

98.33 47.60 99.58 10.25 60.91 33.48 83.38 65.30 76.31 67.56 49.79

9.09 17.54 36.60 6.07 5.79 12.29 9.09 36.03 18.05 16.10 25.52

(7)

Notes: this table presents summary information about our sample countries and firms. Columns 1–2 present for each country the GDP and market capitalization which are measured at December 31, 2010, and are sourced at the World Bank Database. Column 3 presents the mean and the standard deviation of the exchange rate volatilities, which are measured as monthly log difference change in the JP Morgan trade-weighted exchange rate indices, and are sourced at Datastream. The JP Morgan trade-weighted exchange rate index is the number of trade-weighted units of foreign currency per unit domestic currency. Exchange rate arrangements in columns 4–5 are summarized based on the de facto exchange rate arrangement information collected from the IMF Annual Report on Exchange Arrangements and Exchange Restrictions, 1999–2011. Column 6 presents the number of firms in each country, and the number in brackets gives the total number of constituent equities in the major market indices. Column 7 presents the firm-level market capitalization data measured at December 31, 2010, and are sourced at Datastream. The last 3 columns contain the minimum, maximum, and median of zero return ratios as the proportion of daily zero returns out of all daily return observations.

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Countries with non-float regimes China 5926.61 4762.84 Morocco 90.80 69.15

Firm-level data FX volatility-mean/ std. dev. (%)

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

163

board), [b] soft pegs (conventional peg, pegged exchange rates within horizontal bands, crawling pegs, stabilized arrangements, and crawl-like arrangements), [c] floating regimes (floating and free floating), and [d] other managed arrangements (AREAER) (2011: 1). Descriptions of each category can be seen in Appendix B, and Appendix C contains detail on the de facto exchange rate arrangements during our sample period — December 31, 1999 to December 31, 2010. Of the four broad categories, our sample countries have had soft pegs, other managed arrangements, and floating arrangements during the sample period. We collapse these categories into two: pegged (soft pegs and other managed arrangements) and floating (floating arrangements).4 The summarized information based on our non-float/floating classification is reported in columns 4 and 5 of Table 1. Two countries – China and Morocco – had exclusively non-floating exchange rate arrangements during the period. For seven countries – Egypt, Hungary, Indonesia, Malaysia, Peru, Russia, and Turkey – there were periods of non-floating and floating exchange rate regimes during 1999 to 2010, and the remainder – Brazil, Chile, Colombia, Czech Republic, India, Korea, Mexico, the Philippines, Poland, South Africa, and Thailand – had a floating exchange rate throughout the period.5 We use the world stock market index from Morgan Stanley Capital International, and the emerging market index is from Standard & Poor's.6 We use the JP Morgan monthly trade-weighted indexes as our exchange rate measure, and an increase in the rate indicates an appreciation of the currency. These data were obtained from Datastream. The means and standard deviations of the exchange rate volatilities for each country are reported in column 3 of Table 1. As expected, countries with exclusively non-floating regimes have smaller exchange rate volatilities than countries that experienced regime variation over time, as well as those with floating exchange rate systems. The number of firms in each country is reported in column 6 of Table 1. Our sample comprises the constituent firms of each country's major equity market index (these indexes are listed in Appendix A); this yields a sample of 2007 firms. As noted by Bekaert et al. (2007),7 thin trading is a well-known problem with emerging market financial data. Following Bekaert et al. (2007), we calculate the proportion of zero returns out of all daily return observations for each firm; summary statistics on these can be seen in columns 8, 9 and 10. The firm with the highest proportion of zero returns has 99.58% zero (Corferias from Colombia). In two markets – Indonesia and Peru – the median zero return is greater than 50%. Given that the high proportion of zero returns may affect our regression estimates, we exclude firms that are severely thinly-traded; specifically, firms that have more than 30% zero returns, or have no trading data for 22 consecutive days (one month). As an additional filter, we remove firms that have less than 252 (one year's) observations, so as to ensure sufficient data for the regression analysis. After further removing financial firms8 our final data set comprises 1523 firms. The number of firms in each country varies from 7 in the Czech Republic to 307 in Korea. As can be seen in column 7 of Table 1, there is substantial variation in firm size across countries, as measured by market capitalization in the US dollars as at 31 December 2010. Russia has the largest firms, with a median value of US$4011 million, and Hungary has by far the smallest at US$67 million.

4 For the four markets that had ‘other managed arrangements’ during the sample period, ‘other managed arrangements’ appeared only for a very short time. We include the period when ‘other managed arrangement’ was in place to make our time series data consecutive and thus to facilitate our analysis. Further, because ‘other managed arrangement’ is characterized by frequent shifts in policies (see in Appendix B) and is closer to non-floating regimes, we include it in the category of non-floating. 5 We use the IMF de facto exchange rate regime classification, which is available from 1999. As the IMF regime data cover different periods for each country, to ensure comparability between countries our data period starts on December 31, 1999 (some countries' exchange rate information is not available before this) and ends on December 31, 2010. 6 The correlation between world market returns and emerging market returns is 0.67. We examined the possibility of multicollinearity problems when using both of these stock indexes. The results of diagnostic tests proposed by Belsley et al. (1980) suggest that there are no significant multicollinearity issues. 7 Bekaert et al. (2007) examined the liquidity of a set of emerging market firms similar to those we consider here, and found that the proportion of daily zero returns out of all daily return observations was as high as 52% (in Colombia); the average proportion of daily zero returns across the 19 emerging markets was found to be 31%. 8 This is because the exchange rate exposure of financial firms will be different from that of non-financial firms due to their different asset and liability structures as well as access to hedging instruments (Chue and Cook, 2008).

164

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Table 2 Summary of the foreign exchange exposure estimates. Country

No. of firms

(1) Non-floating regimes China 297 Morocco 28 Average Non-floating and floating regimes Egypt 11 Hungary 24 Indonesia 85 Malaysia 17 Peru 9 Russia 40 Turkey 58 Average Floating regimes Brazil 53 Colombia 15 Czech 7 Chile 26 India 152 Korea 307 28 Mexico Philippines 11 Poland 226 South Africa 94 Thailand 35 Average

Significant (at 10% or better)

Mean

No. (%)

+/−

All

Pos.

Neg.

Pos. & sig.

Neg. & sig.

(2)

(3)

(4)

(5)

(6)

(7)

(8)

130 (44%) 8 (29%)

0/130 3/5

−0.287 −0.082 −0.185

0.080 0.488 0.284

−0.315 −0.509 −0.412

/ 1.098 1.098

−0.439 −0.894 −0.667

5 (45%) 14 (58%) 58 (68%) 7 (41%) 3 (33%) 16 (40%) 56 (97%)

1/4 14/0 58/0 7/0 0/3 14/2 56/0

−0.073 0.125 0.423 0.100 −0.292 0.182 0.545 0.144

0.178 0.186 0.481 0.132 0.098 0.306 0.545 0.275

−0.167 −0.547 −0.341 −0.049 −0.404 −0.245 / −0.292

0.428 0.252 0.566 0.211 / 0.499 0.564 0.420

−0.244 / / / −0.861 −0.472 / −0.526

43 (81%) 8 (53%) 5 (71%) 10 (38%) 134 (88%) 113 (37%) 17 (61%) 7 (64%) 71 (31%) 49 (52%) 11 (31%)

40/3 8/0 1/4 4/6 134/0 76/37 13/4 7/0 54/17 2/47 9/2

0.239 0.122 −0.283 0.006 0.432 0.041 0.130 0.309 0.064 −0.142 0.080 0.091

0.277 0.135 0.083 0.077 0.460 0.135 0.203 0.386 0.152 0.037 0.147 0.190

−0.124 −0.054 −0.430 −0.090 −0.584 −0.085 −0.088 −0.035 −0.120 −0.185 −0.088 −0.171

0.318 0.198 0.164 0.182 0.491 0.226 0.291 0.467 0.276 0.060 0.255 0.266

−0.145 / −0.519 −0.137 / −0.171 −0.117 / −0.294 −0.278 −0.226 −0.431

Notes: this table presents the results from estimation of Eq. (5) on 1523 firms from 20 emerging markets for the period December 31, 1999 to December 31, 2010. For each country, we present the number of firms in the sample in column 1, the number and proportion of corresponding exchange rate exposure coefficients (βXn) which are statistically significant in column 2, the number of significantly negative and positive exchange rate exposure coefficients in column 3. Columns 4–8 give the average value of foreign exchange exposures of all firms, positively exposed firms, negatively exposed firms, significantly positively exposed firms and significantly negatively exposed firms, respectively, for each country.

3.3. Foreign exchange exposure estimates Table 2 summarizes the results from estimation of Eq. (5) for all firms by country. It is clear that the frequency of significant exchange rate exposure and the magnitude and direction of exposures vary markedly across countries and firms. As presented in column 2, the percentage of firms whose equity returns are significantly affected by exchange rate changes ranges from nearly all for Turkey (56 out of 58 firms) to Morocco where around 30% (8 out of 28) of firms are significantly exposed. Overall, half of the sample firms (765 out of 1523) have significant exchange rate exposure; this is markedly higher than the rates of exposure found in other studies of firms in emerging markets (Aysun and Guldi, 2011; Chue and Cook, 2008; Kiymaz, 2003; Lin, 2011). As can be seen in column 3, firms in Turkey, Brazil, Colombia, Hungary, India, Indonesia, Korea, Malaysia, Mexico, the Philippines, Poland, Russia and Thailand tend to be positively exposed; that is, appreciating currencies are associated with rising stock prices. In contrast, firms in China, Morocco, Egypt, Peru, Czech, Chile and South Africa are more likely to be negatively exposed. Differences in the direction of exchange rate change on equity prices may be related to firms' net asset-liability exposure (Lin, 2011), exporting and importing activities, or the extent of import competition that they face. In this study, we are not particularly concerned about the direction of exchange rate effects on equity returns, but rather the

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

165

extent of exposure in an absolute sense. However, we frequently disaggregate exchange rate effects by sign of the exposure coefficient in order to preserve information; over-aggregating can obscure interesting effects. Table 2 also includes summary information on the average value of foreign exchange exposure of all firms (column 4), positively exposed firms (column 5), negatively exposed firms (column 6), significantly positively exposed firms (column 7) and significantly negatively exposed firms (column 8) for each country. It is clear that firms in non-floating regimes suffer the highest foreign exchange exposure as measured by mean absolute exposure coefficients. For example, the mean absolute exposure for firms in non-float regimes is 1.10 for positively exposed firms and 0.67 for negatively exposed firms. The equivalent means for firms in countries with both floating and non-float regimes are 0.42 and 0.53 respectively, and for firms in countries with exclusively floating exchange rate regimes 0.27 and 0.43 respectively. These findings suggest that, consistent with Parsley and Popper (2006) non-float regimes are indeed associated with greater foreign exchange exposure than flexible regimes. 3.4. Exchange rate exposure in sub-periods Their economies have also experienced positive changes in the last decade and a half, such as deepening financial reform and greater participation by firms in international business. They have also faced major challenges, particularly those associated with the 2008/2009 financial crisis. Because these episodes may have affected the nature and extent of emerging market firms' foreign exchange exposure experience, as well as for the purpose of examining the robustness of our exposure estimates, we examine firm-level exposure during four sub-periods. Given the widely documented evidence that the stock market rationally signals changes in real economic activity,9 the sub-periods are based on the time-varying performance of global and emerging markets. The four sub-periods are depicted in Fig. 1. The first (SUB1) is from December 31, 1999 to Dec 31, 2002 — the immediate aftermath of the emerging market crises of the late 1990s (Chue and Cook, 2008).10 The second sub-period (SUB2) starts from January 1, 2003 and ends on December 31, 2007 – the pre-global crisis period. The third (SUB3) covers period January 1, 2008 to December 31, 2008 – the ongoing crisis period; and the fourth (SUB4) is from January 1, 2009 to December 31, 2010 – the immediate post-crisis period. Following Parsley and Popper (2006), we add interactive time dummies to Eq. (5) to capture period-by-period exposure: X4 X W E β D X þ εn;t Rn;t ¼ α n þ βn Rw;t þ βn Re;t þ i¼4 i;n i;t t qffiffiffiffiffiffiffiffiffi 2 With ε n;t ¼ μ n;t σ n;t Xp Xq 2 2 2 and σ n;t ¼ γ þ τσ þ φε i¼1 i n;t−i j¼1 j n;t− j

ð6Þ

where Di,t is the time dummy variable: D1 = 1, December 31, 1999 to Dec 31, 2002; = 0, otherwise; D2 = 1, January 1, 2003 to December 31, 2007; = 0, otherwise; D3 = 1, January 1, 2008 to December 31, 2008; = 0, otherwise; and D4 = 1, January 1, 2009 to December 31, 2010; = 0, otherwise. Table 3 presents the estimation results. As well as the original data filters described in Section 3.2, firms that have less than 22 (one month) observations for the first sub-period are removed to ensure sufficient data for our period-by-period analysis. The sample is therefore reduced to 1058 firms. Columns 2, 4, 6 and 8 present the proportion of firms significantly exposed in each period. Consistent with our full-period results, high levels of exposure are apparent across all four sub-periods, though the percentage of significantly exposed firms varies among periods. Specifically, the proportion of firms with significant exposure fell from SUB1 to SUB2 in most countries, with the average decreasing from 40% to 34%. This reduction in exposure is consistent with Chue and Cook (2008), who found that exchange rate exposure was lower for the period 2002–2006 relative to 1999–2002. The recent financial crisis saw an 9 Several empirical studies have found that stock market prices tend to fluctuate/co-integrate with aggregate economic variables (see for example Cheung and Ng, 1998; Gan et al., 2006; Hosseini et al., 2011; Kwon and Shin, 1999). 10 In Chue and Cook (2008), this period ended on June 30th, 2002. Given the performance of markets as depicted in Fig. 1, we suggest that a more appropriate end date is December 31, 2002.

166

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

1,800 1,600 1,400

MSCI World Market Index

1,200 1,000 800 600

S&P Emerging Market Index 400 200 0 2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Fig. 1. The MSCI World Market Index and S&P Emerging Market Index. This figure depicts the MSCI World Market Index and S&P Emerging Market Index for the period December 31, 1999 to December 31, 2010. All data are from Datastream.

increase in the proportion of firms significantly exposed, with the average rising to 40% during SUB3. Exposure appears to have broadened in SUB4, with the average rising again to 46% during SUB4. Columns 3, 5, 7 and 9 present the mean absolute exposure for the significantly exposed firms in each period. The greatest exposure occurred during SUB1 — immediately after the emerging market crises of the late 1990s. Average exposure fell in most countries from SUB1 to SUB2, from 1.017 to 0.558. Similarly, the recent crisis saw an increase in the magnitude of exchange rate exposure, with the average rising to 0.861 during SUB3, and then decreasing slightly to 0.715 during SUB4. Overall, we find foreign exchange exposure is pervasive among emerging market firms across all of the sub-periods. Though the severity of foreign exchange exposure varies over time, we find nothing to suggest that foreign exchange exposure has fallen over the sample period — in prevalence or extent. It is clear that the financial crisis (period SUB3) brought with it higher exchange rate exposure for emerging market firms; there was a substantial increase in the proportion of firms exposed as well as the magnitude of significant exposure. While more firms were “infected” with foreign exchange exposure in the immediate post-crisis period (as suggested by the increasing proportion of significant exposure from SUB3 to SUB4), the reduction in average exposure suggests that concerns about emerging market firms' foreign exchange exposure waned as the worst of the crisis passed. Alternatively, it is possible that during the crisis firms became much more aware of the risks associated with currency movements, and in response improved and expanded their exchange rate risk management activities.

4. How do changes in exchange rate regime affect foreign exchange exposure? 4.1. Model specification Among our 20 emerging markets, seven countries experienced both floating and non-floating regimes during the research period. This allows us to examine whether and how firm-level exchange rate exposure

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

167

Table 3 Summary of the period-by-period foreign exchange exposure. Country

No. of firms

SUB1

SUB2

SUB3

SUB4

No. sig. (%)

Mean sig.

No. sig (%)

Mean sig.

No. sig (%)

Mean sig.

No. sig (%)

Mean sig.

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

15.25 9.09

0.656 0.065

4.04 27.27

0.579 0.925

15.25 27.27

1.114 3.513

80.72 9.09

1.259 2.334

Non-floating and floating regimes Egypt 6 50.00 Hungary 17 29.41 Indonesia 56 50.00 Malaysia 14 42.86 Peru 5 60.00 Russia 12 66.66 Turkey 50 98.00

2.367 0.927 0.632 0.606 3.798 1.003 0.500

50.00 23.53 67.86 7.14 40.00 66.67 98.00

0.191 0.368 1.054 0.227 1.307 0.638 0.700

16.67 58.82 30.36 78.57 60.00 16.67 90.00

0.577 0.389 3.310 0.927 0.708 0.678 0.667

50.00 41.17 46.43 21.43 20.00 33.34 70.00

0.636 0.338 1.249 0.234 1.259 0.382 0.721

Floating regimes Brazil 32 Chile 24 Colombia 10 Czech 4 India 108 Korea 267 Mexico 23 Philippines 6 Poland 87 South Africa 74 Thailand 29 Average

0.883 0.231 2.423 0.321 1.525 0.646 0.655 0.733 0.565 0.610 1.185 1.017

81.25 33.33 90.00 25.00 65.74 19.48 39.13 16.67 28.74 47.30 27.59 34.22

0.478 0.182 0.362 0.546 0.714 0.412 0.252 0.944 0.621 0.317 0.350 0.558

37.50 41.66 20.00 0.00 62.04 46.07 56.52 66.67 28.74 48.65 10.35 39.79

0.292 0.327 0.178 / 0.795 0.328 0.458 0.682 0.581 0.364 0.478 0.861

25.00 33.34 20.00 25.00 67.60 28.83 26.09 0.00 29.89 25.67 20.69 45.94

0.339 0.400 1.204 0.300 0.596 0.375 0.456 / 0.383 0.284 0.831 0.715

(1) Non-floating regimes China 223 Morocco 11

75.00 45.83 60.00 50.00 28.71 43.45 65.21 100.00 22.99 44.59 58.62 39.51

Notes: this table presents the results from estimation of Eq. (6) on a sample of 1058 firms from 20 emerging markets for the four sub-periods, which are selected with respect to global and emerging equity market performance (see Fig. 1). For each country, we present the number of firms in the sample in column 1, and the percentage and the mean absolute value of significant exchange rate exposure coefficients for each period in columns 2–9, respectively. The first sub-period (SUB1) is from December 31, 1999 to December 31, 2002 — the immediate aftermath of the emerging market crises of the late 1990s; the second (SUB2) is from January 1, 2003 to December 31, 2007 — the pre-crisis (the recent global crisis) period; the third (SUB3) covers the period January 1, 2008 to December 31, 2008 — the ongoing crisis period; and the fourth (SUB4) is from January 1, 2009 to December 31, 2010 — the immediate post-crisis period.

alters when the exchange rate regimes change. Following Parsley and Popper (2006), we add an interactive regime dummy to Eq. (5) to capture regime effect on firms' exposure:

W

E

X

D

Rn;t ¼ α n þ βn Rw;t þ βn Re;t þ βn X t þ βn Dn;t X t þ εn;t qffiffiffiffiffiffiffiffiffi With ε n;t ¼ μ n;t σ 2n;t Xp Xq 2 2 2 and σ n;t ¼ γ þ τσ þ φε i¼1 i n;t−i j¼1 j n;t− j

ð7Þ

where Dn,t equals 1 when a non-floating regime is in place for firm n at period t; and is 0 otherwise. The key parameter of interest is βD n , which measures how the effect of exchange rate movements on stock return alters when exchange rate regime changes. βXn and (βXn + βD n ) provide estimates of foreign exchange exposure for firm n under float and non-float regimes, respectively. 4.2. Empirical results Table 4 presents the estimates of foreign exchange exposure and the interactive regime dummy from estimation of Eq. (7) on the 186 firms from the seven emerging markets that experienced both non-floating

168

Table 4 Individual market tests of the estimates of foreign exchange exposure and the interactive regime dummy.

Egypt Hungary Indonesia Malaysia Peru Russia Turkey Total

No. of firms

βXn sig. at 10% or better

βD n sig. at 10% or better

Aver. sig. βD n

Aver. all |βXn|

No. (%)

+/−

No. (%)

+/−

(1)

(2)

(3)

(4)

(5)

5 18 85 15 9 12 42 186

2 (40%) 9 (50%) 58 (68%) 9 (60%) 4 (44%) 5 (42%) 42 (100%) 129 (69%)

1/1 9/0 58/0 8/1 0/4 5/0 42/0 123/6

1 (20%) 4 (22%) 10 (12%) 3 (20%) 2 (22%) 2 (17%) 8 (19%) 30 (16%)

0/1 0/4 1/9 2/1 2/0 2/0 7/1 14/16

|βXn

+

βD n|

No. (%)

No. (%)

(Float)

(Non-float)

(|βXn| b |βXn + βD n |)

(6)

(7)

(8)

(9)

(10)

(11)

0.162 0.176 0.500 0.164 0.454 0.222 0.571 0.321

0.148 0.111 1.176 0.156 0.511 0.956 1.371 0.633

2 (40%) 7 (39%) 56 (66%) 7 (47%) 5 (56%) 10 (83%) 34 (81%) 121 (65%)

0.069 0.333 0.540* 0.073* 0.822 0.252 0.563 0.379

0.178 0.070 1.795 0.226 0.771 0.780 2.799 0.946

1 (100%) 4 (100%) 9 (90%) 3 (100%) 1 (50%) 1 (50%) 8 (100%) 27 (90%)

b

+

βD n |)

|βXn + βD n|

(Non-float)

(Float)

|βXn

|βXn|

(|βXn|

Notes: this table presents the estimates of foreign exchange exposure and the interactive regime dummy from estimation of Eq. (7) on 186 firms from the 7 emerging markets which experienced both non-floating and floating exchange rate regimes, for the full period from December 31, 1999 to December 31, 2010. For each country, we present the number of firms in the sample in column 1, the number and proportion (in bracket) of significant exchange rate exposure coefficients βXn in column 2, the number of significantly negative and positive exchange rate exposure coefficients in column 3, the number and proportion of significant interactive term coefficients βD n in column 4, the number of significantly negative and positive interactive term exchange rate exposure coefficients in column 5. Columns 6 and 7 (9 and 10) gives the average absolute value of foreign exchange exposures of all (significantly exposed) firms, when float and non-float regimes are in place, respectively. Column 8 (11) gives the number and proportion (in bracket) of all (significantly exposed) firms which have larger foreign exchange exposure under non-float regimes relative to float regimes, in terms of the absolute value of foreign exchange exposure. * indicates that the average value of |βXn| is calculated by assuming that βXn equals to 0 for those 3 firms in Indonesia and X Malaysia which have significant βD n while insignificant βn .

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Country

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

169

and floating exchange rate regimes during our data period.11 For each country, we present the number of firms in the sample in column 1, the number and proportion of significant exchange rate exposure coefficients βXn (column 2), the number of significantly negative and positive exchange rate exposure coefficients (column 3), the number and proportion of significant interaction term coefficients (βD n ) in column 4, and the number of significantly negative and positive interactive term exchange rate exposure coefficients (column 5). Columns 6 and 7 give the average absolute value of foreign exchange exposures of the sample when float and non-float regimes are in place, respectively; and columns 9 and 10 do the same for the significantly exposed firms. Columns 8 and 11 provide the number and proportion of firms that have larger absolute foreign exchange exposure under non-float regimes relative to float regimes. As can be seen in columns 2–5, a high proportion (129 out of 186 or 69%) of firms have significant, mostly positive, exchange rate exposure. Under non-float regimes slightly more firms are exposed (132 or 71%),12 also mostly positively. Regarding the magnitude of the exposure, the results in columns 6–11 suggest that absolute X exposure is in most cases greater under non-float regimes than floating regimes (|βXn + βD n | N |βn|). This is particularly clear when examining firms that are significantly exposed (columns 9 and 10). For example, Indonesian firms that are significantly exposed during both float and non-float periods experience much greater exposure when the non-float regime was in place (an average of 1.795) compared to during floating periods (an average of 0.540). The difference in absolute exposure between the two regimes is frequently statistically significant. Specifically, among the firms that exhibit a change in exposure in non-float regimes relative to floating regimes, 90% have larger exposure when the non-float regime is in place.13 Patnaik and Shah (2010) also found that their sample of 100 Indian firms had greater exposure in periods when exchange rates were less flexible — within a broad framework of a dollar peg. Parsley and Popper (2006) compared the proportion of firms that are significantly exposed under different regimes, and found that in some cases more firms were significantly exposed to the dollar with a peg than without one. Our results suggest that non-float regimes are associated with not only higher proportion of significantly exposed firms, but also larger magnitude of firms' exposure. In the following session, we further explore the question — whether the regime factor is a potential determinant of the magnitude of firms' foreign exchange exposure. 5. Can the exchange rate regime explain firm-level foreign exchange exposure? 5.1. Model specification In this section, we investigate whether exchange rate regime contributes to an explanation of the overall exchange rate exposure of emerging market firms. We conduct a pooled cross-sectional regression, whereby the exchange rate exposure coefficient (estimated in Eq. (5)) is regressed on a regime dummy variable as well as several of firm-level and country-level control variables. As the firm- and country-level traits can assist only in explaining the magnitude rather than the direction of the exchange rate exposure (Aggarwal and Harper, 2010), we take the absolute values of the exchange rate exposure coefficients, |βXn|. The estimation model is as follows:   Xi F Xj C  X γ F þ γ C βn  ¼ α n þ ωReg Regn þ 1 i i 1 j j

ð8Þ

where βXn is the exchange rate exposure coefficient estimated from Eq. (5). Regn is the regime dummy, which equals 1 when a non-float regime is in place and 0 when there is a floating regime. Fi and Cj are firmand country-level control variables, respectively. 5.1.1. Firm-specific variables We control for firm size as proxied by market value (MV), as it is well established that large firms are less likely to be affected by foreign exchange exposure than small firms (Bodnar and Wong, 2003; Chow et 11

For this analysis we remove firms that have less than 22 (one month) observations in either regime period. X D X X D 27 out of 30 firms with significant βD n have significant βn ; the rest 3 have only significant βn while insignificant βn . As (βn + βn ) measures firms' foreign exchange exposure under non-float regimes, the number of firms with significant exposure under non-float regimes should be 132 (=129 + 3). 13 Only in two markets (Hungary and Peru) do we find a reduction in exchange rate exposure associated with a non-floating exchange rate regime. Specifically, all 4 Hungarian firms and 1 out of 2 Peruvian firms have smaller exposure under the non-float regime. 12

170

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

al., 1997; Dominguez and Tesar, 2006; Hutson and Stevenson, 2010), because large firms are more likely to be internationally diversified and thus ‘naturally’ hedged (Aggarwal and Harper, 2010) and because hedging activity is associated with economies of scale (Allayannis and Ofek, 2001; Hagelin and Pramborg, 2006). We include the debt-to-assets ratio (DA) which is a proxy for expected financial distress; He and Ng (1998) argue that firms with higher leverage are more likely to hedge, and they find an inverse relation between exposure and leverage. Market to book ratio (MTBV), which is a common proxy for growth opportunities, is included because hedging theory suggests that high market-to-book firms are more likely to hedge and therefore experience lower exchange exposure (Hutson and Stevenson, 2010). We also include asset turnover (AT) — a proxy for the ability to adjust to changing foreign exchange exposures. As argued by Aggarwal and Harper (2010), firms with better asset management as measured by higher asset turnover should have a more natural protection against changes in pricing in competitive environments and should exhibit lower exposures to foreign exchange risk. As noted by Chue and Cook (2008), an exchange rate movement is a combination of a shock and a monetary-policy response, and exchange rate depreciation may be avoided when the domestic monetary authority responds to an external shock by raising the interest rate. As the ability of market makers to provide liquidity to domestic markets depends on the local cost of funds (Brunnermeier and Pedersen, 2009; Chordia et al., 2005; Coughenour and Saad, 2004), illiquid shares that depend more on market makers may be hurt by an increase in the domestic interest rate, and benefit from the alternative — an exchange rate depreciation. We therefore include turnover ratio (TR), 14 measured by the number of shares traded for a stock on a particular day divided by the total number of shares outstanding, to proxy a stock's liquidity. We also include the quick ratio (QR), as a measure of short-term firm-level liquidity. Nance et al. (1993) suggested that firms can reduce the likelihood of incurring financial distress by holding liquid assets. Liquidity can therefore be seen as a substitute for hedging; firms with high levels of short-term liquidity are therefore likely to be more exposed (Hutson and Stevenson, 2010). Finally, we include proxies for foreign institutional ownership and government institutional ownership — free float foreign held (FH) and free float government held (GH) respectively. ‘Free float foreign held’ is calculated as the percentage of total shares on issue held by a government or government institution, and ‘free float government held’ is the percentage of total shares in issue held by a foreign institution. These different types of owners may induce the firm to pursue alternative hedging strategies, and may also react to exchange rate movements differently (Chue and Cook, 2008).

5.1.2. Country-specific variables Our country-level explanatory variables include trade openness, calculated by exports plus imports as a percentage of GDP (EIG). Hutson and Stevenson (2010) found a strongly robust inverse relation between the trade openness and firm-level foreign exchange exposure. It is well documented that a higher per capita GDP is related to a better-developed local bond market (Burger and Warnock, 2007), which translates to less reliance on foreign currency debt (Chue and Cook, 2008) and thus lower foreign exchange exposure. We use GDP (PCG) as a proxy for a country's overall level of development. The final country-level variable is a measure of the country's financial development. Better-developed financial markets are associated with lower hedging costs and a wider range of available hedging tools, both of which can increase firms' hedging activities and thus reduce their foreign exchange exposure (Chue and Cook, 2008). Following Chue and Cook (2008), we use the ratio of broad money (the sum of narrow money and quasi-money) to GDP (BMG) to proxy country-level financial development.15

14 Turnover ratio by value used in Chue and Cook (2008) is not available for most of our emerging market firms. We use turnover by volume instead. 15 We also use another measure of financial development — stock market capitalization as a percentage of GDP, proposed by Garcia and Liu (1999), and the results are qualitatively the same. We also considered some other country-level variables, such as international reserves to external debt ratio and external debt to GDP ratio, suggested by Chue and Cook (2008). However, data for these two variables are not available for many of our sample countries.

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

171

5.2. Data and variable descriptions The summary statistics for our firm- and country-level control variables are presented in Appendix D. For DA, MTBV, AT, TR and QR, we take an average for each firm over the period of interest; for MV, the natural log of its average value is used because market value is highly right-skewed (the skewness is 12.90). To calculate the mean value, annual data are used. For some financial statement variables there are extreme values. We therefore winsorize the data by setting the highest and lowest 1% of these variables to their levels at precisely 1% and 99%, respectively. For three country-level variables – EIG, PCG and BMG – we take averages of their annual figures over the 11 years from 2000 to 2010.

5.3. Pooled cross-sectional inter-market analysis Table 5 presents the results for our pooled cross-sectional regression analysis of Eq. (8) on the subset of data comprising firms from the 13 countries that had exclusively floating or exclusively non-floating exchange rate regimes during the full sample period. In Panel A we report the findings for the full sample. As can be seen, the exchange rate regime dummy (REG) is positive but not significant. As the inclusion of firms that are not significantly exposed creates noise in the data set, we follow Aggarwal and Harper (2010) and Hutson and O'Driscoll (2010) and re-estimate Eq. (8) with subsamples that include firms that are significantly exposed (at the 0.10 level) (Panel B), significantly positively exposed (Panel C), and significantly negatively exposed (Panel D). The REG dummy is significantly positive (at the 1% level) in all three of these estimations. After controlling for many firm- and country-level factors known to affect firm-level exposure, we find that firm-level foreign exchange exposure is also affected by exchange rate regime factor, and is significantly higher when there is a non-float regime than under floating regimes. Table 5 Summary results of the pooled cross-sectional inter-market analysis. Panel A: full sample (1)

(2)

(3)

Coeff.

Std. error 0.056 0.008 0.044 0.003 0.009 0.001 0.006 0.001 0.001 0.000 0.000 0.001 0.046

REG 0.066 MV −0.006 DA 0.108 MTBV 0.001 AT −0.024 TR 0.004 QR −0.003 GH −0.001 FH 0.0002 EIG −0.001 PCG 0.0001 BMG −0.001 Const. 0.483 Adj. R-sq. 0.257 N 1254

Panel B: significant

Panel C: positive and significant

Panel D: negative and significant

(4)

(7)

(10)

(5)

(6)

p-Value Coeff.

Std. error

0.24 0.48 0.01 0.84 0.01 0.00 0.66 0.32 0.99 0.01 0.00 0.14 0.00

0.091 0.013 0.061 0.003 0.015 0.001 0.011 0.001 0.001 0.001 0.000 0.001 0.077

0.248 −0.028 0.154 0.001 −0.013 0.005 0.015 −0.001 0.0004 0.001 0.0002 −0.003 0.646 0.300 583

(8)

(9)

p-Value Coeff.

Std. error

0.01 0.04 0.01 0.69 0.39 0.00 0.19 0.11 0.91 0.14 0.00 0.01 0.00

0.250 0.016 0.067 0.003 0.018 0.002 0.013 0.001 0.001 0.001 0.000 0.001 0.074

0.657 −0.044 0.148 −0.004 −0.007 0.003 0.009 −0.001 0.0003 −0.001 0.0001 0.001 0.605 0.351 345

(11)

(12)

p-Value Coeff.

Std. error

p-Value

0.01 0.01 0.03 0.17 0.70 0.05 0.50 0.19 0.66 0.07 0.00 0.35 0.00

0.168 0.026 0.146 0.010 0.021 0.002 0.023 0.007 0.002 0.001 0.000 0.002 0.171

0.00 0.55 0.96 0.29 0.36 0.01 0.84 0.08 0.32 0.03 0.01 0.01 0.01

0.484 0.015 0.008 0.011 −0.019 0.005 0.005 −0.013 0.002 −0.003 0.0001 −0.005 0.454 0.379 238

Notes: this table presents the results for our pooled cross-sectional regression analysis of Eq. (8) on the 627 firms from the 13 emerging markets where the exchange rate regime was either float or non-float, for the full period from December 31, 1999 to December 31, 2010. REG denotes the regime dummy variable, which takes a value of 1 in non-floating regimes and 0 otherwise. MV denotes market value; DA is short for debt to assets ratio; MTBV is the market to book ratio; AT denotes asset turnover; TR denotes turnover ratio; QR denotes quick ratio; GH and FH are the percentage of total shares in issue held by a government (or government institution) and foreign institution, respectively; EIG is the ratio of exports plus imports as a percentage of GDP; PCG denotes per capita GDP; and BMG is the ratio of broad money to GDP. We present the coefficient, standard error and significant probability of regime dummy coefficient (ωReg) for all the 627 firms in Panel A (columns 1–3), for significantly exposed firms in Panel B (columns 4–6), for positively significantly exposed firms in Panel C (columns 7–9) and for negatively significantly exposed firms in Panel D (columns 10–12), respectively. All standard errors are heteroskedasticity-consistent.

172

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Consistent with prior literature (Aggarwal and Harper, 2010; Chue and Cook, 2008; Hutson and Stevenson, 2010), the market value (MV) coefficient is negatively significant, confirming the inverse relation between exchange exposure and firm size. However, the coefficients on the other two financial variables – debt-to-assets ratio (DA) and turnover ratio (TR) – do not have the predicted negative sign. The positive sign for debt-to-assets implies that firms with a greater likelihood of financial distress (as proxied by leverage) are more exposed to exchange rate movements. This is plausible if firms with high leverage do not fully hedge (Chow and Chen, 1998), because adverse shocks to exchange rates increase the likelihood of financial distress owing to high interest payment commitments (Hutson and Stevenson, 2010). The positive and significant turnover ratio suggests that more liquid firms have larger exposure. The predicted negative and significant coefficient for asset turnover (AT) only appears when estimating the full sample (Panel A). Hedging theories suggest that firms with high market-to-book ratio (MTBV) are more likely to hedge and therefore have lower exchange exposure (Hutson and Stevenson, 2010). However, insignificant market-to-book ratio suggests that this theory may not be applicable to emerging market firms; firms with high market-to-book ratio may not hedge, or may only partially hedge. The findings on the quick ratio (QR) are consistent with empirical results of Hutson and Stevenson (2010), who also found an insignificant coefficient for quick ratio. For the final two firm-specific variables – free float foreign held (FH) and free float government held (GH) – the insignificant coefficients suggest that foreign and government ownerships do not affect the degree of exchange rate exposure. This may be explained by the fact that there are relatively low levels of foreign institutional ownership (an average of 1.09%) and government ownership (5.37%) among our sample firms. At the country level, we find that a higher trade openness (EIG) tends to have a negative effect on exchange rate exposure, which is in contrast with the findings of Hutson and Stevenson (2010), who examined developed country firms. The coefficient on per capita GDP (PCG) has the predicted positive sign and is highly significant, implying that a better-developed local bond market reduces the reliance on foreign-currency debt and thus reduces foreign exchange exposure. The expected negative coefficient for the ratio of broad money to GDP (BMG) appears when estimating significantly exposed (Panel B) and negatively significantly exposed firms (Panel D). As a better-developed financial markets are associated with lower hedging costs and a wider range of available hedging tools, this finding is consistent with well-developed financial markets being associated with greater hedging activity. 5.4. Pooled cross-sectional intra-market analysis Table 6 presents the results for the pooled cross-sectional regression analysis of Eq. (8) on the subset of data consisting firms from the seven countries that had float and non-float exchange rate regimes during the sample period. (This is the same data set as used in the estimations reported in Table 4). In this analysis, βXn equals βXn when a floating regime is in place and (βXn + βD n ) when there is a non-float regime, estimated from Eq. (7). Given that there are only a small number of firms (261 − 237 = 24) with significant negative exposure estimates, we are unable to include an analysis with the negatively exposed firms (as we did in Table 5 Panel D). Across all of the estimations, the regime coefficient (REG) is highly significantly positive. The results reported in Table 6 confirm that the exchange rate regime is a very important determinant of firm-level exchange rate exposure for emerging market firms.16 Apparently stable exchange rate arrangements fail to protect firms against exposure to exchange rate movements. Non-float exchange rate regimes in fact appear to exacerbate firm-level foreign exchange exposure. There are several possible explanations for our findings. In standard pegged systems, the peg stabilizes only one bilateral exchange rate, while the others often remain volatile (Parsley and Popper, 2006). Our findings support the moral hazard hypothesis, which suggests that pegged exchange rates lull firms into engaging in unhedged foreign cash flow or balance sheet transactions that leave them exposed to sudden changes in their 16 A number of commonalities are found across Tables 5 and 6 in the firm-level and the country-level control variable coefficients. The exceptions are debt-to-asset ratio (DA) and asset turnover (AT) that are found to be not significant in any specification, and turnover ratio (TR) is only significant in the full sample analysis. At the country level, per capita GDP (PCG) is not significant and broad money to GDP (BMG) is significant only for the positively exposed group.

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

173

Table 6 Summary results of pooled cross-sectional intra-market analysis. Panel A: full sample (1)

REG MV DA MTBV AT TR QR GH FH EIG PCG BMG Const. Adj. R-sq. N

(2)

Panel B: significant (3)

(4)

(5)

Panel C: positive and significant (6)

(7)

(8)

(9)

Coeff.

Std. error

p-Value

Coeff.

Std. error

p-Value

Coeff.

Std. error

p-Value

0.431 −0.194 0.229 −0.006 0.037 0.013 0.067 −0.001 0.006 −0.007 0.0001 0.007 0.820 0.112 372

0.099 0.084 0.465 0.023 0.081 0.008 0.059 0.004 0.008 0.003 0.000 0.005 0.207

0.00 0.02 0.62 0.79 0.65 0.09 0.26 0.80 0.40 0.03 0.97 0.15 0.00

0.463 −0.331 −0.009 −0.084 0.078 0.010 0.054 0.000 0.015 −0.011 0.000 0.014 1.184 0.149 261

0.119 0.153 0.032 0.565 0.113 0.008 0.044 0.006 0.013 0.005 0.000 0.009 0.361

0.00 0.03 0.78 0.88 0.49 0.22 0.22 1.00 0.25 0.03 0.92 0.10 0.00

0.452 −0.307 −0.011 0.145 0.122 0.009 0.036 0.002 0.021 −0.016 0.000 0.023 0.735 0.169 237

0.134 0.152 0.038 0.584 0.113 0.008 0.044 0.006 0.015 0.006 0.000 0.010 0.341

0.00 0.04 0.78 0.80 0.28 0.29 0.42 0.75 0.16 0.00 0.30 0.02 0.03

Notes: this table presents the results for the pooled cross-sectional regression analysis of Eq. (8) on the 186 firms from the 7 emerging markets which experienced both non-floating and floating exchange rate regimes, for the full period from December 31, 1999 to December 31, 2010. REG denotes the regime dummy variable; MV denotes market value; DA is short for debt to assets ratio; MTBV is the market to book ratio; AT denotes asset turnover; TR denotes turnover ratio; QR denotes quick ratio; GH and FH are the percentage of total shares in issue held by a government (or government institution) and foreign institution, respectively; EIG is the ratio of exports plus imports as a percentage of GDP; PCG denotes per capita GDP; and BMG is the ratio of broad money to GDP. We present the coefficient, standard error and significant probability of regime dummy coefficient (ωReg) for all the 186 firms in Panel A (columns 1–3), for significantly exposed firms in Panel B (columns 4–6), and for positively significantly exposed firms in Panel C (columns 7–9), respectively. Given that there is only a small number (261 − 237 = 24) of negatively significant exposure estimates, we are unable to do the similar estimation as reported in Table 5 Panel D.

currency's value as a result of external shocks. A related explanation is that under non-floating exchange rate regimes, firms (or their creditors) may be less aware of exchange rate risk, and thus do not actively engage in hedging to mitigate their foreign exchange exposure (Kamil, 2006, 2012). Firms and their creditors may well be ignorant or insensible of exchange rate risks in pegged systems, but our findings suggest that the market is highly attuned to these risks. At the heart of the foreign exchange exposure of firms in non-floating regimes is the unsustainability of pegged exchange rate pegged systems. Klein and Shambaugh (2008) found that while there is tremendous variability in the duration of pegs, they tend to last for a median of only 2 years, and only 28% of pegs last more than five years. The sample of countries examined in this section included those in which pegged exchange rate regimes are rather short-lived. Even in markets in which ex-post exchange rate arrangements did not alter during the sample period (those included in the analysis in Section 5.3 above), investors may have been anticipating that the regime will change at some time in the future. In such environments, business decisions involving foreign currency transactions would be very difficult to properly assess, and firms would also face challenges in making foreign exchange risk management decisions. 6. Summary and conclusions In this paper, we have investigated the effects of exchange rate regimes on firm-level foreign exchange exposure. Using a data set of 1523 firms from 20 emerging markets for the period December 1999 to December 2010, we have examined three questions. First, how exposed are these emerging market firms? Using market-based measures of firm-level foreign exchange exposure, we find that around half of our sample firms exhibit significant exposure to exchange rate movements, much higher than those found in studies of developed market firms (see Bartram and Bodnar, 2007, for a review). Our sub-period exposure analysis shows that these findings are robust to changes in economic conditions, including the recent global financial crisis. By categorizing our sample markets into three groups – two countries with

174

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

exclusively pegged exchange rate arrangements, 11 with exclusively floating exchange rates, and 7 that experienced both types of regime during our sample period – we find that the magnitude of exchange rate exposure is larger for firms in countries with pegged exchange rates than for those in countries with floating arrangements. Second, how do firm-level exchange rate exposures change when the exchange rate regime changes? The 7 markets which experienced both floating and non-floating exchange rate regimes during the sample period provide us with a rich data set to test this issue. Our results suggest that the switch to a non-floating exchange rate regime is associated with more widespread exchange rate exposure, and more substantially a larger magnitude of firm-level exposure. Third, to what extent can exchange rate regimes explain the magnitude of firm-level foreign exchange exposure? Our cross-sectional analysis has yielded strong and consistent evidence that the non-float exchange rate regime heightens firm-level exchange rate risk, and this holds after controlling for a wide range of firm-level and country-level exposure determinants. Our findings provide insight into the ‘fear of floating’ debate. Non-floating regimes fail to protect firms from exchange rate exposure, and may even exacerbate exposure. Exchange rate regime matters at the micro as well as the macro level. The ‘moral hazard hypothesis’ advanced by Eichengreen and Hausmann (1999) could play an important role in explaining our findings. Pegged exchange rates may well promote risk-taking and diminish firms' incentives to hedge; they may also contribute to an environment in which there is less awareness of exchange rate risk (Kamil, 2006, 2012). Our findings imply that emerging market company managers should more carefully monitor and manage their foreign exchange risks, and policymakers should be more cognizant of the dangers that firms face when operating in highly managed exchange rate regimes.

Acknowledgments We would like to thank discussants and participants of the 2013 FMA International Conference (USA), 2013 FMA Asian Conference (China) and the Chinese Economists Society (CES) Conference (China). We are also grateful to our colleagues, Prof. John Cotter and Prof. Don Bredin, at Michael Smurfit Business School for their valuable comments and suggestions on earlier drafts of the paper. All errors are our responsibility.

Appendix A. The major equity market index for each country Country

Equity market index

Datastream mnemonic

Brazil Chile China Colombia Czech Egypt Hungary India Indonesia Malaysia Mexico Morocco Peru Philippines Poland Russia South Africa South Korea Thailand Turkey

BRAZIL BOVESPA CHILE SELECTIVE (IPSA) SHANGHAI SE COMPOSITE COLOMBIA-DS Market PRAGUE SE PX EGYPT EGX 30 HUNGARY-DS Market INDIA BSE (200) NATIONAL IDX COMPOSITE FTSE BURSA MALAYSIA KLCI MEXICO IPC (BOLSA) MOROCCO MOST ACTIVE PERU-DS NON-FINANCIAL PHILIPPINE SE I(PSEi) WARSAW GENERAL INDEX RUSSIA RTS INDEX FTSE/JSE ALL SHARE KOREA SE COMPOSITE (KOSPI) BANGKOK S.E.T. 50 ISTANBUL SE NATIONAL 100

BRBOVES IPSASEL CHSCOMP TOTMKCB CZPXIDX EGCSE30 TOTMKHN IBOMBSE JAKCOMP FBMKLCI MXIPC35 MADEX TOTLIPE PSECOMP POLWIGI RSRTSIN JSEOVER KORCOMP BNGK50 TRKISTB

Notes: this appendix presents for each country the major equity market index that is applied in this paper. All data are collected from Datastream.

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

175

Appendix B. Descriptions of IMF de facto exchange rate arrangements Category

Sub-category

Descriptions

Hard pegs

No separate legal tender

Classification as an exchange rate arrangement with no separate legal tender involves confirmation of the country authorities' de jure exchange rate arrangement. The currency of another country circulates as the sole legal tender (formal dollarization). Adopting such an arrangement implies complete surrender of the monetary authorities' control over domestic monetary policy. Classification as a currency board involves confirmation of the country authorities' de jure exchange rate arrangement. A currency board arrangement is a monetary arrangement based on an explicit legislative commitment to exchange domestic currency for a specified foreign currency at a fixed exchange rate, combined with restrictions on the issuance authority to ensure the fulfillment of its legal obligation. This implies that domestic currency is usually fully backed by foreign assets, eliminating traditional central bank functions such as monetary control and lender of last resort, and leaving little room for discretionary monetary policy. Some flexibility may still be afforded, depending on the strictness of the banking rules of the currency board arrangement. Classification as a conventional peg involves confirmation of the country authorities' de jure exchange rate arrangement. For this category, the country formally (de jure) pegs its currency at a fixed rate to another currency or a basket of currencies, where the basket is formed, for example, from the currencies of major trading or financial partners and weights reflect the geographic distribution of trade, services, or capital flows. The anchor currency or basket weights are public or notified to the IMF. The country authorities stand ready to maintain the fixed parity through direct intervention (i.e., via sale or purchase of foreign exchange in the market) or indirect intervention (e.g., via exchange-rate-related use of interest rate policy, imposition of foreign exchange regulations, exercise of moral suasion that constrains foreign exchange activity, or intervention by other public institutions). There is no commitment to irrevocably keep the parity, but the formal arrangement must be confirmed empirically: the exchange rate may fluctuate within narrow margins of less than ±1% around a central rate — or the maximum and minimum values of the spot market exchange rate must remain within a narrow margin of 2% for at least six months. Classification as a stabilized arrangement entails a spot market exchange rate that remains within a margin of 2% for six months or more (with the exception of a specified number of outliers or step adjustments) and is not floating. The required margin of stability can be met with respect to either a single currency or a basket of currencies, where the anchor currency or the basket is ascertained or confirmed using statistical techniques. Classification as a stabilized arrangement requires that the statistical criteria are met and that the exchange rate remains stable as a result of official action (including structural market rigidities). The classification does not imply a policy commitment on the part of the country authorities. Classification as a pegged exchange rate within horizontal bands involves confirmation of the country authorities' de jure exchange rate arrangement. The value of the currency is maintained within certain margins of fluctuation of at least ±1% around a fixed central rate, or a margin between the maximum and minimum value of the exchange rate that exceeds 2%. It includes arrangements of countries in the ERM of the European Monetary System, which was replaced with the ERM II on January 1, 1999, for countries with margins of fluctuation wider than ±1%. The central rate and width of the band are public or notified to the IMF. Classification as a crawling peg involves confirmation of the country authorities' de jure exchange rate arrangement. The currency is adjusted in small amounts at a fixed rate or in response to changes in selected quantitative indicators, such as past inflation differentials vis-à-vis major trading partners or differentials between the inflation target and expected inflation in major trading partners. The rate of crawl can be set to generate inflation-adjusted changes in the exchange rate (backward looking) or set at a predetermined fixed rate and/or below the projected inflation differentials

Currency board

Soft pegs

Conventional peg

Stabilized arrangement

Pegged exchange rate within horizontal bands

Crawling peg

176

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Appendix (continued) B (continued) Category

Sub-category

Soft pegs

Crawl-like arrangement

Floating arrangements

Floating

Free floating

Other managed arrangements

Other managed arrangements

Descriptions (forward looking). The rules and parameters of the arrangement are public or notified to the IMF. For classification as a crawl-like arrangement, the exchange rate must remain within a narrow margin of 2% relative to a statistically identified trend for six months or more (with the exception of a specified number of outliers), and the exchange rate arrangement cannot be considered as floating. Usually, a minimum rate of change greater than allowed under a stabilized (peg-like) arrangement is required; however, an arrangement is considered crawl-like with an annualized rate of change of at least 1%, provided the exchange rate appreciates or depreciates in a sufficiently monotonic and continuous manner. A floating exchange rate is largely market determined, without an ascertainable or predictable path for the rate. In particular, an exchange rate that satisfies the statistical criteria for a stabilized or a crawl-like arrangement is classified as such unless it is clear that the stability of the exchange rate is not the result of official actions. Foreign exchange market intervention may be either direct or indirect and serves to moderate the rate of change and prevent undue fluctuations in the exchange rate, but policies targeting a specific level of the exchange rate are incompatible with floating. Indicators for managing the rate are broadly judgmental (e.g., balance of payments position, international reserves, parallel market developments). Floating arrangements may exhibit more or less exchange rate volatility, depending on the size of the shocks affecting the economy. A floating exchange rate can be classified as free floating if intervention occurs only exceptionally and aims to address disorderly market conditions and if the authorities have provided information or data confirming that intervention has been limited to at most three instances in the previous six months, each lasting no more than three business days. If the information or data required are not available to the IMF staff, the arrangement is classified as floating. Detailed data on intervention or official foreign exchange transactions will not be requested routinely of member countries — only when other information available to the staff is not sufficient to resolve uncertainties about the appropriate classification. This category is a residual and is used when the exchange rate arrangement does not meet the criteria for any of the other categories. Arrangements characterized by frequent shifts in policies may fall into this category.

Notes: this appendix gives descriptions of IMF de facto exchange rate arrangements. The information about classifications and descriptions of exchange rate arrangements are collected from IMF Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER) 2011, Pg. 50–52.

Country

1999

Brazil Chile

Floating (as of Dec 31) Free floating (as of Dec 31) Conventional peg (as of Dec 31)

China

Colombia Czech Republic Egypt

Other managed arrangement (as of Dec 31)

India

Free floating (as of Dec 31) Free floating (as of Dec 31)

2001

2002

2003

2004 2005

2006

Crawl-like arrangement (Aug 1)

Free floating (as of Dec 31) Floating (as of Dec 31) Conventional peg (as of Dec 31)

Hungary

Indonesia

2000

2007 2008

2009

Stabilized arrangement (June 1)

2010

Crawl-like arrangement (June 21)

Floating (Jan 1) Free floating (June 30) Pegged exchange rate within horizontal bands (Jan 30) Pegged exchange rate within horizontal bands (October 1)

Floating (January 1)

Free Floating (Jan 1) Floating (Jan 29)

Conventional peg (Feb 1)

Other managed Crawl-like arrangement arrangement (Aug 1) (Mar 12)

Free floating (Feb 26)

Floating (Mar 1)

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Appendix C. Summary of de facto exchange arrangements reported in IMF AREAER (December 31, 1999–December 31, 2010)

Floating (Jan 1) Floating (June 30)

Stabilized arrangement (June 21)

177

(continued) 1999

South Korea Malaysia

Free floating (as of Dec 31) Conventional peg (as of Dec 31)

Mexico

Free floating (as of Dec 31) Conventional peg (as of Dec 31) Free floating (as of Dec 31)

Morocco Peru

South Africa Turkey Thailand

Floating (as of Dec 31) Crawling peg (as of Dec 31) Free floating (as of Dec 31)

2001

2002

2003

2004 2005

2006

2007 2008

Floating (Jan 1)

Floating (Jan 1)

Floating (Dec 31)

2010

Crawl-like arrangement (Sep 16)

Stabilized arrangement (Mar 1)

Free floating (Feb 22) Floating (June 30)

2009 Floating (May 1) Other managed arrangement (Jan 1) Floating (Oct 9)

Other managed arrangement (Nov 1)

Floating (Oct 24)

Free Floating (Aug 4)

Floating (Oct 4)

Notes: this appendix gives a summary of de facto exchange rate arrangements of our sample countries over the period from December 31, 1999–December 31, 2010. According to IMF definition, exchange rate arrangements are categorized into: [a] hard pegs (exchange arrangements with no separate legal tender and currency board), [b] soft pegs (conventional peg, pegged exchange rates within horizontal bands, crawling pegs, stabilized arrangements, and crawl-like arrangements), [c] floating arrangements (floating and free floating), and [d] other managed arrangements (see IMF

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Philippines Floating (as of Dec 31) Poland Free floating (as of Dec 31) Russia Free floating (as of Dec 31)

2000

178

Country

Country

Firm-level data

Country-level Data

Market value (US$ m)

Debt to assets

Market to book

Asset turnover

Turnover ratio

Quick ratio

Government held (%)

Foreign held (%)

(EX + IM) to GDP

Per capita GDP

Broad money to GDP

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

0.33 0.31 0.28 0.14 0.30 0.26 0.17 0.26 0.33 0.26 0.28 0.22 0.15 0.20 0.36 0.18 0.19 0.18

3.76 1.95 3.81 1.32 1.94 3.11 1.62 4.50 2.19 1.17 3.44 2.31 2.81 6.71 2.13 1.97 2.20 3.16

0.66 0.66 0.68 0.45 0.94 0.63 0.98 0.89 0.95 1.07 0.65 0.82 0.83 0.77 0.49 1.34 0.83 1.36

36.74 0.38 4.40 1.14 0.65 3.66 0.79 3.01 0.87 7.08 0.67 0.89 0.47 1.18 0.34 2.21 0.28 0.58

1.24 1.37 1.03 1.31 1.04 1.66 4.85 1.15 1.64 1.21 1.62 1.65 1.57 1.23 1.03 1.39 2.02 1.25

2.87 0.00 0.02 1.77 5.32 13.12 0.69 1.22 2.37 0.05 4.96 0.00 0.74 0.00 1.01 1.32 6.00 1.38

8.31 7.57 0.62 4.14 39.40 1.15 21.17 4.96 11.97 2.61 2.10 3.06 11.24 21.93 6.78 9.07 14.17 4.04

25.50 71.91 57.36 35.59 139.42 53.55 146.81 38.81 58.26 81.40 198.27 57.38 69.67 42.36 91.21 74.19 56.50 58.69

8911.79 12,265.31 4507.81 7546.62 20,404.55 4977.73 16,340.33 2262.29 3308.83 23,006.76 11,730.31 12,494.11 3684.50 6822.41 2972.06 14,167.91 11,994.24 8674.47

52.65 76.72 143.62 29.39 63.62 84.01 50.90 61.99 41.76 65.92 127.19 27.03 87.02 30.11 56.16 44.17 31.13 65.89

0.35 0.19

1.77 2.24

1.14 1.16

1.12 13.15

0.97 3.42

2.08 0.40

3.10 2.32

134.10 48.59

6813.65 10,698.70

109.43 38.71

179

Panel A: summary information Brazil 6841.71 Chile 2956.44 China 1990.03 Colombia 5028.47 Czech 3774.45 Egypt 2411.85 Hungary 709.61 India 3784.51 Indonesia 807.91 Korea 471.36 Malaysia 4779.19 Mexico 4933.75 Morocco 735.56 Peru 1425.16 Philippines 1984.04 Poland 526.72 Russia 14,372.39 South 4378.51 Africa Thailand 1979.30 Turkey 1105.57

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Appendix D. Summary information and statistics of the firm- and country-level variables

Debt to assets

Market to book

Asset turnover

Turnover ratio

Quick ratio

Government held (%)

Foreign held (%)

(EX + IM) to GDP

Per capita GDP

Broad money to GDP

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

2.59 1.92 15.01 0.07 2.57 12.07 1498

0.96 0.82 3.96 0.00 1.89 8.37 1516

3.40 1.31 40.72 0.00 3.93 20.49 1523

1.25 0.99 6.86 0.14 3.12 15.40 1515

1.09 0.00 99.00 0.00 8.34 91.07 1520

5.37 0.00 95.00 0.00 3.51 16.80 1520

67.03 58.69 198.27 25.50 2.12 10.48 1523

10,915.18 8911.79 23,006.76 2262.29 0.54 1.94 1523

74.47 65.89 143.62 27.03 0.98 2.51 1523

Panel B: summary statistics Mean 2282.85 0.25 Median 395.73 0.24 Maximum 195,988.40 0.63 Minimum 2.35 0.00 Skewness 12.90 0.34 Kurtosis 257.35 2.40 Obs. 1523 1516

180

Market value (US$ m)

Panel C: correlation matrix MV

DA

MTBV

AT

TR

QR

GH

FH

EIG

PCG

BMG

REG*

1.00 0.07 0.30 −0.21 −0.19 −0.03 0.23 0.06 −0.24 −0.40 0.14 0.16

1.00 −0.01 −0.22 0.05 −0.40 −0.02 −0.05 0.00 −0.07 0.15 0.00

1.00 0.02 −0.03 −0.02 0.02 0.06 −0.24 −0.41 0.23 0.35

1.00 −0.03 −0.14 −0.01 0.05 0.09 0.17 −0.22 −0.27

1.00 −0.01 −0.05 −0.13 −0.03 0.19 0.05 0.23

1.00 0.02 0.04 0.01 0.03 −0.14 −0.05

1.00 −0.03 0.01 −0.05 −0.08 −0.07

1.00 0.05 0.00 −0.22 −0.23

1.00 0.46 0.06 −0.34

1.00 −0.36 −0.31

1.00 0.42

1.00

Notes: this table presents summary data on our firm- and country-level variables. Panel A shows for each country the mean values of the 8 firm-level variables (columns 1–8) and the 3 country-level variables (columns 9–11). Market value (MV) is the share price multiplied by the number of ordinary shares in issue. Debt to assets (DA) is the ratio of total debt to total assets. Market to book (MTBV) is the ratio of the market value of the ordinary (common) equity to the balance sheet value of the ordinary (common) equity in the company. Asset turnover (AT) is the ratio of net sales or revenues to total assets. Turnover ratio (TR) is the number of shares traded for a stock on a particular day divided by the total number of ordinary shares that represent the capital of the company. Quick ratio (QR) is calculated as the sum of cash (and equivalents) and receivables divided by current liabilities. Government (GH)/Foreign held (FH) denotes the percentage of total shares in issue held by a government (or government institution)/foreign institution. All these firm level data are collected from Datastream. (EX + IM) to GDP (EIG) is the ratio of exports plus imports as a percentage of GDP, and is drawn from the World Bank Database. Per Capita GDP (PCG) denotes the gross domestic product divided by the number of people in the country, and is collected from the IMF World Economic Outlook Database. Broad Money to GDP (BMG) is the ratio of the sum of narrow money and quasi-money as a percentage of GDP, and is drawn from the World Bank Database. Panel B presents the summary statistics of the above firm- and country-level data series. Panel C presents the correlation matrix for these 11 control variables and the independent variable, the regime dummy (REG), which equals 1 when the firm n is from a country that had at any time during the data period a non-float phase, and equals 0 when the country had a floating regime for the whole period. For REG (indicated by *), we calculate the Spearman rank-order correlations, and for other variables, we calculate the Pearson correlations.

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

MV DA MTBV AT TR QR GH FH EIG PCG BMG REG*

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

181

References Aggarwal, R., Harper, J.T., 2010. Foreign exchange exposure of “domestic” corporations. J. Int. Money Financ. 29 (8), 1619–1636. Aggarwal, R., Inclan, C., Leal, R., 1999. Volatility in emerging stock markets. J. Financ. Quant. Anal. 34 (1), 33–55. Akaike, H., 1973. Information theory and an extension of the maximum likelihood principle. In: Petrov, B.N., Caski, F. (Eds.), Second International Symposium on Information Theory. Springer-Verlag, pp. 267–281. Allayannis, G., Ofek, E., 2001. Exchange rate exposure, hedging, and the use of foreign currency derivatives. J. Int. Money Financ. 20 (2), 273–296. Amihud, Y., 1994. Exchange rates and the valuation of equity shares. In: Amihud, Y., Levich, R.M. (Eds.), Exchange Rates and Corporate Performance. Irwin, New York, pp. 49–59. Arteta, C.O., 2005. Exchange rate regimes and financial dollarization: does flexibility reduce bank currency mismatches? Berkeley Electron. J. Macroecon., Top. Macroecon. 5 (1), 1226–1246. Aysun, U., Guldi, M., 2011. Derivatives market activity in emerging markets and exchange rate exposure. Emerg. Mark. Financ. Trade 47 (6), 46–67. Baek, S.-G., 2013. On the determinants of aggregate currency mismatch. J. Policy Model 35 (4), 623–637. Baillie, R.T., Bollerslev, T., 1989. The message in daily exchange rates: a conditional-variance tale. J. Bus. Econ. Stat. 7 (3), 297–305. Bartov, E., Bodnar, G.M., 1994. Firm valuation, earnings expectations, and the exchange-rate exposure effect. J. Financ. 49 (5), 1755–1785. Bartram, S.M., 2004. Linear and nonlinear foreign exchange rate exposures of German nonfinancial corporations. J. Int. Money Financ. 23 (4), 673–699. Bartram, S.M., Bodnar, G.M., 2007. The exchange rate exposure puzzle. Manag. Financ. 33 (9), 642–666. Bartram, S.M., Karolyi, G.A., 2006. The impact of the introduction of the Euro on foreign exchange rate risk exposures. J. Empir. Financ. 13 (4–5), 519–549. Bekaert, G., 1995. Market integration and investment barriers in emerging equity markets. World Bank Econ. Rev. 9 (1), 75–107. Bekaert, G., Harvey, C.R., 2002. Research in emerging markets finance: looking to the future. Emerg. Mark. Rev. 3 (4), 429–448. Bekaert, G., Harvey, C.R., Lundblad, C., 2007. Liquidity and expected returns: lessons from emerging markets. Rev. Financ. Stud. 20 (6), 1783–1831. Belsley, D.A., Kuh, E., Welsch, R.E., 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. Wiley, New York. Bemdt, E.K., Hall, B.H., Hall, R.E., Hausman, J.A., 1974. Estimation and inference in nonlinear structural models. Ann. Econ. Soc. Meas. 3 (4), 653–665. Bodnar, G.M., Wong, M., 2000. Estimating exchange rate exposures: some “ weighty” issues. NBER Working Paper No. 7497 http://dx. doi.org/10.3386/w7497. Bodnar, G.M., Wong, M.F., 2003. Estimating exchange rate exposures: issues in model structure. Financ. Manag. 32 (1), 35–67. Bollerslev, T., Chou, R.Y., Kroner, K.F., 1992. ARCH modeling in finance. J. Econom. 52 (1), 5–59. Brunnermeier, M.K., Pedersen, L.H., 2009. Market liquidity and funding liquidity. Rev. Financ. Stud. 22 (6), 2201–2238. Burger, J.D., Warnock, F.E., 2007. Foreign participation in local currency bond markets. Rev. Financ. Econ. 16 (3), 291–304. Burnside, C., Eichenbaum, M., Rebelo, S., 2001. Hedging and financial fragility in fixed exchange rate regimes. Eur. Econ. Rev. 45 (7), 1151–1193. Calvo, G.A., Reinhart, C.M., 2002. Fear of floating. Q. J. Econ. 117 (2), 379–408. Carrieri, F., Errunza, V., Hogan, K., 2007. Characterizing world market integration through time. J. Financ. Quant. Anal. 42 (4), 915. Chang, R., Velasco, A., 2000. Financial fragility and the exchange rate regime. J. Econ. Theory 92 (1), 1–34. Cheung, Y.-W., Ng, L.K., 1998. International evidence on the stock market and aggregate economic activity. J. Empir. Financ. 5 (3), 281–296. Choi, J.J., Jiang, C., 2009. Does multinationality matter? Implications of operational hedging for the exchange risk exposure. J. Bank. Financ. 33 (11), 1973–1982. Chordia, T., Sarkar, A., Subrahmanyam, A., 2005. An empirical analysis of stock and bond market liquidity. Rev. Financ. Stud. 18 (1), 85–129. Chow, E.H., Chen, H.-L., 1998. The determinants of foreign exchange rate exposure: evidence on Japanese firms. Pac. Basin Financ. J. 6 (1), 153–174. Chow, E.H., Lee, W.Y., Solt, M.E., 1997. The exchange-rate risk exposure of asset returns. J. Bus. 70 (1), 105–123. Chue, T.K., Cook, D., 2008. Emerging market exchange rate exposure. J. Bank. Financ. 32 (7), 1349–1362. Coughenour, J.F., Saad, M.M., 2004. Common market makers and commonality in liquidity. J. Financ. Econ. 73 (1), 37–69. Cruz-Rodriguez, A., 2013. Choosing and assessing exchange rate regimes: a survey of the literature. Econ. Anal. Rev. 28 (2), 37–61. Doidge, C., Griffin, J., Williamson, R., 2006. Measuring the economic importance of exchange rate exposure. J. Empir. Financ. 13 (4–5), 550–576. Dominguez, K.M.E., Tesar, L.L., 2001. A reexamination of exchange-rate exposure. Am. Econ. Rev. 91 (2), 396–399. Dominguez, K.M.E., Tesar, L.L., 2006. Exchange rate exposure. J. Int. Econ. 68 (1), 188–218. Eichengreen, B., Hausmann, R., 1999. Exchange Rates and Financial Fragility http://dx.doi.org/10.3386/w7418. Engle, R.F., 1982. Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation. Econometrica 50 (4), 987–1007. Fischer, S., 2001. Distinguished lecture on economics in government: exchange rate regimes: is the bipolar view correct? J. Econ. Perspect. 15 (2), 3–24. Flood, J.E., Lessard, D.R., 1986. On the measurement of operating exposure to exchange rates: a conceptual approach. Financ. Manag. 15 (1), 25–36. Frankel, J.A., Rose, A.K., 2005. Is trade good or bad for the environment? Sorting out the causality. Rev. Econ. Stat. 87 (1), 85–91. Gan, C., Lee, M., Yong, H.H.A., Zhang, J., 2006. Macroeconomic variables and stock market interactions: New Zealand evidence. Invest. Manag. Financ. Innov. 3 (4), 89–101. Garcia, V.F., Liu, L., 1999. Macroeconomic determinants of stock market development. J. Appl. Econ. 2 (1), 29–59. Ghosh, A., Gulde, A.-M., Wolf, H., 2002. Exchange rate regimes: classification and consequences. In: Ghosh, A., Gulde, A.-M., Wolf, H. (Eds.), Exchange Rate Regimes: Choices and Consequences. MIT Press, Cambridge.

182

M. Ye et al. / Emerging Markets Review 21 (2014) 156–182

Hagelin, N., Pramborg, B., 2006. Empirical evidence concerning incentives to hedge transaction and translation exposures. J. Multinatl. Financ. Manag. 16 (2), 142–159. He, J., Ng, L.K., 1998. The foreign exchange exposure of Japanese multinational corporations. J. Financ. 53 (2), 733–753. Hekman, C.R., 1983. Measuring foreign exchange exposure: a practical theory and its application. Financ. Anal. J. 39 (5), 59–65. Hosseini, S.M., Ahmad, Z., Lai, Y.W., 2011. The role of macroeconomic variables on stock market index in China and India. Int. J. Econ. Financ. 3 (6), 233–243. Hsieh, D.A., 1989. Modeling heteroscedasticity in daily foreign-exchange rates. J. Bus. Econ. Stat. 7 (3), 307–317. Hutson, E., O'Driscoll, A., 2010. Firm-level exchange rate exposure in the Eurozone. Int. Bus. Rev. 19 (5), 468–478. Hutson, E., Stevenson, S., 2010. Openness, hedging incentives and foreign exchange exposure: a firm-level multi-country study. J. Int. Bus. Stud. 41 (1), 105–122. International Monetary Fund (IMF), 2011. Annual Report on Exchange Arrangements and Exchange Restrictions. International Monetary Fund, Washington, DC. Jorion, P., 1990. The exchange-rate exposure of U.S. multinationals. J. Bus. 63 (3), 331. Kamil, H., 2006. Does moving to a flexible exchange rate regime reduce currency mismatches in firms' balance sheets? Paper presented at the 7th Jacques Polak Annual Research Conference (November 9–10, 2006). Washington, DC. Kamil, H., 2012. How do exchange rate regimes affect firms' incentives to hedge currency risk? Micro Evidence for Latin America. IMF Working Paper No. 12/69 (http://ssrn.com/abstract=2028245). Kiymaz, H., 2003. Estimation of foreign exchange exposure: an emerging market application. J. Multinatl. Financ. Manag. 13 (1), 71–84. Klein, M.W., Shambaugh, J.C., 2006. Fixed exchange rates and trade. J. Int. Econ. 70 (2), 359–383. Klein, M.W., Shambaugh, J.C., 2008. The dynamics of exchange rate regimes: fixes, floats, and flips. J. Int. Econ. 75 (1), 70–92. Korajczyk, R.A., 1996. A measure of stock market integration for developed and emerging markets. World Bank Econ. Rev. 10 (2), 267–289. Kwon, C.S., Shin, T.S., 1999. Cointegration and causality between macroeconomic variables and stock market returns. Glob. Financ. J. 10 (1), 71–81. Lahiri, A., Végh, C.A., 2002. Living with the fear of floating: an optimal policy perspective. In: Edwards, S., Frankel, J.A. (Eds.), Preventing Currency Crises in Emerging Markets. University of Chicago Press, pp. 663–704. Levy-Yeyati, E., Sturzenegger, F., 2003. To float or to fix: evidence on the impact of exchange rate regimes on growth. Am. Econ. Rev. 93 (4), 1173–1193. Lin, C.H., 2011. Exchange rate exposure in the Asian emerging markets. J. Multinatl. Financ. Manag. 21 (4), 224–238. Loudon, G., 1993. The foreign exchange operating exposure of Australian stocks. Account. Financ. 33 (1), 19–32. Marston, R.C., 2001. The effects of industry structure on economic exposure. J. Int. Money Financ. 20 (2), 149–164. Martínez, L., Werner, A., 2002. The exchange rate regime and the currency composition of corporate debt: the Mexican experience. J. Dev. Econ. 69 (2), 315–334. McKinnon, R.I., 2000. The East Asian dollar standard, life after death? Econ. Notes 29 (1), 31–82. Mishkin, F.S., 1997. Understanding financial crises: a developing country perspective. NBER Working Paper No. 5600 http://dx.doi. org/10.3386/w5600. Muller, A., Verschoor, W.F., 2006a. Foreign exchange risk exposure: survey and suggestions. J. Multinatl. Financ. Manag. 16 (4), 385–410. Muller, A., Verschoor, W.F.C., 2006b. Asymmetric foreign exchange risk exposure: evidence from U.S. multinational firms. J. Empir. Financ. 13 (4–5), 495–518. Nance, D.R., Smith, C.W., Smithson, C.W., 1993. On the determinants of corporate hedging. J. Financ. 48 (1), 267–284. Nydahl, S., 1999. Exchange rate exposure, foreign involvement and currency hedging of firms: some Swedish evidence. Eur. Financ. Manag. 5 (2), 241–257. Parsley, D.C., Popper, H.A., 2006. Exchange rate pegs and foreign exchange exposure in East and South East Asia. J. Int. Money Financ. 25 (6), 992–1009. Patnaik, I., Shah, A., 2010. Does the currency regime shape unhedged currency exposure? J. Int. Money Financ. 29 (5), 760–769. Petreski, M., 2009. Exchange-rate regime and economic growth: a review of the theoretical and empirical literature. Economics Discussion Paper No. 2009-31 http://dx.doi.org/10.2139/ssrn.1726732. Rossi, J.J.L., 2009. Corporate financial policies and the exchange rate regime: evidence from Brazil. Emerg. Mark. Rev. 10 (4), 279–295. Rossi, J.J.L., 2011. Exchange rate exposure, foreign currency debt, and the use of derivatives: evidence from Brazil. Emerg. Mark. Financ. Trade 47 (1), 67–89. Schneider, M., Tornell, A., 2004. Balance sheet effects, bailout guarantees and financial crises. Rev. Econ. Stud. 71 (3), 883–913. Schwarz, G., 1978. Estimating the dimension of a model. Ann. Stat. 6 (2), 461–464. Shin, H.-H., Soenen, L., 1999. Exposure to currency risk by US multinational corporations. J. Multinatl. Financ. Manag. 9 (2), 195–207. Tse, Y., 1998. International transmission of information: evidence from the Euroyen and Eurodollar futures markets. J. Int. Money Financ. 17 (6), 909–929.