مواجهه با نرخ ارز و نوسانات بازده بخشی: مدارک و شواهد از بخش های صنعتی ژاپنی
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|8933||2008||22 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 9680 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Japan and the World Economy, Volume 20, Issue 4, December 2008, Pages 639–660
Most studies of exchange rate exposure of stock returns do not address three relevant aspects simultaneously. They are, namely: sensitivity of stock returns to exchange rate changes; sensitivity of volatility of stock returns to volatility of changes in foreign exchange market; and the correlation between volatilities of stock returns and exchange rate changes. In this paper, we employ a bivariate GJR-GARCH model to examine all such aspects of exchange rate exposure of sectoral indexes in Japanese industries. Based on a sample data of fourteen sectors, we find significant evidence of exposed returns and its asymmetric conditional volatility of exchange rate exposure. In addition, returns in many sectors are correlated with those of exchange rate changes. We also find support for the “averaged-out exposure and asymmetries” argument. Our findings have direct implications for practitioners in formulating investment decisions and currency hedging strategies.
Modeling exchange rate exposure has been an important growing area of research in the last decade. The literature of exchange rate exposure dates back to early 1940s. Initially, a firm's actual rather than future cash flows are used to analyze the exposure. However, such an approach is inappropriate for practical reasons. For instance, realized cash flows do not capture a firm's operating exposure. Changes in exchange rates may also influence the future activities of the firm. And, it is not operationally easy to obtain a significant amount of firm-specific information, especially when the study is focused on a large number of firms. Adler and Dumas (1984) give a lucid review of the definition and measurement of exposure to currency risk. More recently, Bodnar and Wong (2003) provide an excellent account of issues in estimating exchange rate exposures. In this paper we propose a unified approach to address the exchange rate exposure of stock returns. To the best of our knowledge, our approach is the first direct investigation that simultaneously captures three relevant aspects exchange rate exposures including time-varying risk of 14 sectoral returns in Japanese industries. Bivariate GJR-GARCH models are employed to achieve such purposes. We follow the well-documented approach of Adler and Dumas (1984) and others to take a firm's market value as reasonable proxy to its future operating cash flows. Based on the efficient market hypothesis, exchange rate exposure of a firm is defined as “the sensitivity of [its] economic value, or stock price, to exchange rate changes” (Hekman, 1983). Adler and Dumas (1984) show that exchange rate exposure can be obtained by regressing a firm's value on exchange rate. The following augmented market model is often used to estimate exposure coefficients: equation(1) ri,t=δ0,i+δmrm,t+δxrx,t+ξi,t, i=1,2,…,nri,t=δ0,i+δmrm,t+δxrx,t+ξi,t, i=1,2,…,n and ξi,t∼N(0,σ2)ξi,t∼N(0,σ2) Turn MathJax on where ri,t is the returns on firm i's stock at time t; rm,t the returns on market portfolio at time t; rx,t is the changes in exchange rate at time t. Here exchange rate is expressed as local currency price of foreign currency; δm is the firm i's exposure to market returns; δx the firm i's exchange rate exposure coefficient which measures the sensitivity of a firm's returns to the exchange rate movements; ξi,t is the regression residual which is assumed to follow a normal distribution with zero mean and constant variance. Many earlier studies rely heavily on the standard OLS or SUR method of estimation, with emphasis on the sensitivity of stock returns to changes in exchange rate. Among others, such studies include Jorion (1990), Bodnar and Gentry (1993), Chamberlain et al. (1997), Chow and Chen (1998), Dominguez (1998), He and Ng (1998), and Dominguez and Tesar, 2001a, Dominguez and Tesar, 2001b and Dominguez and Tesar, 2006, respectively. Among many studies that focus on Japan, Bodnar and Gentry (1993)1 obtain OLS estimates of monthly exchange rate exposure of industry portfolios,2 and find that 5 out of 20 Japanese industries are significantly exposed to exchange rate changes. They find that an appreciation in the yen affects favourably on both non-traded goods sector producers and importers, and adversely on exporters and the value of their foreign operations. Dominguez (1998) classifies a sample of 275 Japanese firms into 18 portfolios distinguished by industry type, firm size and degree of internationalization. She finds that 7 out of 18 portfolios are significantly exposed to weekly exchange rate changes. At the firm level, He and Ng (1998) find that most of the 171 firm-returns are positively exposed to depreciation of the yen. They observe that significantly exposed firms are mostly concentrated in three sectors: electric machinery; precision instruments; and transport equipment. Contrary to the findings of He and Ng (1998), Chow and Chen (1998) find that Japanese firms are adversely affected by depreciation of the yen. One plausible explanation is that these firms may have anticipated the unavoidable appreciation of the yen and are actually able to respond to it efficiently. The major problem of the augmented market approach is the questionable assumption of time-invariance in the variance of firm's return and in changes of exchange rate. In recent years, it has been common to use generalized autoregressive conditional heteroskedasticity (GARCH)-type models to accommodate the time-varying volatility in empirical studies of exchange rate changes. There are two main categories. The first group employs GARCH-type models to augment the mean equation with a time-varying variance structure in order to improve the precision of parameters. For instances, they include Patro et al. (2002) and Koutmos and Martin (2003a). However, all such studies are confined to either one or two aspects of exchange rate exposure. The second group assigns a more active role to the GARCH structure. For example, Kanas (2000), Apergis and Rezitis (2001) and Yang and Doong (2004) employ bivariate asymmetric GARCH models to analyze the mutual impact of volatilities between equity and exchange rate markets.3Koutmos and Martin (2003b) examine the first- and second-moment exchange rate exposure. However, most of these studies take the country as the unit of analysis. And the main defect of a country study is that exchange rate exposure could be averaged out when a highly aggregated index is used.4 The reason is that various industries or sectors may be exposed negatively/positively to the exchange rate changes depending on whether they are import/export dominant. By the same token, the asymmetries associated with the exchange rate exposure of both first and second moments of stock returns are also likely to be averaged out when highly aggregated indexes are used. On the other hand, there are cases where estimation of overall market's exchange rate exposure may not largely help in hedging and investment decisions. For instance, a local investor who wants to invest in equities which bear low currency risk may find the exposure estimates of industrial sectors much useful, though the overall market's exposure does not appear to be so. As such, there is a need to focus our study on exchange rate exposure of Japanese industries at the sector level. There are several reasons why we think this paper's emphasis on time-varying volatility at the sector level is relevant. First, as volatilities of exchange rate and stock returns are vital for currency hedging and investment decisions, exchange rate exposure of sectoral returns should not be confined to the exposure coefficient (δx) in the mean equation as specified in the augmented market model. Second, there are at least two other aspects of exchange rate exposure of stock returns which are worth investigating: (a) sensitivity of volatility of stock returns to the volatility in foreign exchange market; (b) conditional correlation between stock returns and exchange rate changes. Hence, we intend to fill this gap by engaging a bivariate GJR-GARCH model to capture all three aspects of daily exchange rate exposure simultaneously in 14 Japanese industrial sectors. We find evidence of all three aspects of exchange rate exposure of sectoral returns, thereby supporting our argument that the entire currency risk actually faced by a firm/sector is not fully captured by the exposure coefficient alone in the traditional augmented market model. Moreover, we find evidence to support the “averaged-out exposure” argument, in which exchange rate exposure of sectoral returns can be averaged out when highly aggregated indexes are employed at the market level. We also find support for extending such an argument to asymmetries associated with exchange rate exposure. The rest of this paper is organized as follows. Section 2 briefly describes the datasets used in our study and reports the preliminary analysis of the returns series. Section 3 highlights the gist of the model used to capture the three aspects of exchange exposure of sectoral returns. Section 4 is divided into two parts. The first part reports estimation results and provides some discussions. The second part examines the dynamic properties of exchange rate exposure of returns and their conditional volatility by simulations. We also demonstrate that an indirect effect of the volatility of exchange rate exposure on returns could still be possible even if the sectoral returns are not directly exposed to the exchange rate changes in the mean equation. Some concluding remarks are given in Section 5.
نتیجه گیری انگلیسی
We have employed a bivariate GJR-GARCH model to capture the exchange rate exposure of fourteen Japanese industrial sectors, with emphasis on three aspects of exchange rate exposure: sensitivity of sectoral returns to changes in exchange rate of the yen; sensitivity of the conditional volatility of sectoral returns to that of changes in the exchange rate of the yen and its possibly asymmetric effect; and the correlation between sectoral returns and exchange rate changes. In general, we find strong evidence of exchange rate exposure in all three aspects. This implies that the entire currency risk actually faced by firms is not fully captured by the traditional “exchange rate exposure coefficient” alone. We find that returns on sectors automobile and parts (A&P), electrical and electronic equipment (E&EE), household goods and textiles (HH&T) and information technology and hardware (IT&H) show positive exposure to changes in exchange rate of the yen. And returns in sectors oil and gas (O&G) and construction and building materials (C&BM) are negatively exposed to exchange rate changes. Our results are consistent with the previous studies of Japanese industries at the sector level. In addition, volatility of returns in each of these six sectors is also significantly correlated with that of the exchange rate changes, and the sign of correlation coefficient is largely consistent with that of the exposure coefficient. Moreover, conditional volatility of returns in six sectors including automobile and parts (A&P), diversified industries (DI), electrical and electronic equipment (E&EE), engineering and machinery (E&M), household goods and textiles (HH&T) and software and computer services (S&CS) are positively exposed to that of exchange rate changes, suggesting that volatility in these sectors increases with an increase in volatility of exchange rate changes. Furthermore, we find evidence of asymmetric exposure of the volatility of returns in six sectors (DI, E&EE, E&M, HH&T, O&G and S&CS). In five sectors, volatility of sectoral returns caused by a depreciation of the yen is greater than that caused by an appreciation of the yen of the same magnitude. On the contrary, returns in O&G sector are more vulnerable to appreciation of the yen.16 The simulation exercise reveals some interesting patterns of the dynamics of exchange rate exposure of sectoral returns. First, the impact of an exchange rate shock on returns, though large in magnitude, may die down relatively quickly. Second, even if the returns are not directly exposed to the exchange rate changes, as long as they are sensitive to its own volatility, there could be a persistent indirect impact via the exposure of conditional volatility of the returns to the volatility in foreign exchange markets. Finally, if the volatility of sectoral returns is significantly exposed to the volatility of changes in exchange rate with sufficiently large magnitude, the impact of an exchange rate shock on the conditional volatility of the returns may be even higher than the impact on its own volatility. We also check the empirical validity of the argument that exchange rate exposure of stock returns is averaged out when highly aggregated stock indexes are used. Our findings not only support the “averaged-out exposure” argument, but also provide new evidence that asymmetries associated with exchange rate exposure are also likely to be averaged out with highly aggregated indexes.