بازار ارز و قیمت نفت در آسیا
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|14828||2013||10 صفحه PDF||سفارش دهید||6700 کلمه|
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله شامل 6700 کلمه می باشد.
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Asian Economics, Volume 28, October 2013, Pages 41–50
In this paper, we examine whether oil price can predict exchange rate returns for 14 Asian countries. A new GLS-based time series predictive regression model proposed by Westerlund and Narayan (WN, 2012) is used. The main finding is that higher oil price leads to future depreciation of the Vietnamese dong but future appreciations of the local currencies of Bangladesh, Cambodia, and Hong Kong. A comparison of the widely used Lewellen (2004) and WN (2012) estimators show that both provide similar results in in-sample analysis, although WN is relatively superior at longer horizons in out-of-sample analysis.
Exactly how the price of oil influences exchange rates is well-established in theory. Krugman (1983) and Golub (1983) argue that higher oil prices will transfer wealth from the oil importers to oil exporters. This means that when the oil price rises, oil exporting countries may experience an appreciation of their exchange rate and oil importing countries may see a depreciation of their exchange rate. However, if the oil importer makes up a small share of the oil exporter's export market but a larger share of the oil exporter's imports, then the transfer of wealth from oil importers to oil exporters would improve the oil importer's trade balance, which would mean that the long-run effect on the exchange rate is positive for the oil importers (Corden, 1984 and De Grauwe, 1996). In the last two decades, many studies have provided strong empirical evidence that oil price and exchange rates co-move, and that oil price causes exchange rates; see, inter alia, Wu, Chung, and Chung (2012), Amano and van Norden (1998) and Benhad (2012) for evidence on the US; Mohammadi and Jahan-Parva (2012) for oil exporting countries; Tiwari, Dar, and Bhenja (2013) for evidence on India; Narayan, Narayan, and Prasad (2008) for evidence on Fiji; Akram (2004) for evidence on Norway; Camarero and Tamarit (2002) for evidence on Spain; and Amano and van Norden (1995) for evidence on Canada. Moreover, Chen and Chen (2007) found that oil price predicts exchange rates in G7 countries. Some studies point to weak or no evidence of causation and co-movement. Some examples are Czudaj and Beckmann (2013) and Reboredo (2012) for the EU countries; Huang and Guo (2007) for China; Chen, Rogoff, and Rossi (2010) for Australia, Canada, South Africa, and Chile; Chaudhari and Daniel (1998) for 16 OECD countries; and Chinn (2000) for Asia-Pacific countries. A key feature of this empirical literature is that there are voluminous studies on the short- and long-run co-movement and causation relationships. However, rather surprisingly, not much has been done on the subject of forecasting exchange rates using oil prices. Chen and Chen (2007) and Chen et al. (2010) are the only studies that investigate whether oil prices can predict exchange rates. Chen and Chen (2007) consider the G7 countries and use data from 1972:1 to 2005:10. Using in-sample and out-of- sample forecasting models, they find that real oil price predicts future real exchange rates. They also find that in out-of-sample evaluations, the oil price-based predictor model outperforms the random walk model at various horizons. Chen et al. (2010), on the other hand, contend that commodity prices, including the oil price, are not good predictors of exchange rates. In particular, they find that currencies belonging to four commodity exporting countries (Australia, Canada, South Africa and Chile) cannot be predicted by commodity prices, neither in in-sample nor in out-of-sample evaluations. Thus, it is evident that the limited empirical evidence on exchange rate predictability using the oil price predictor has evinced mixed results. The aim of this paper is to fill the existing empirical research gap on the subject of exchange rate predictability based on the oil price predictor through considering a sample of 14 Asian countries. In terms of methodological contribution, the paper is particularly concerned with the selection of estimators when conducting in-sample and out-of-sample forecasting of real exchange rate returns by using the real oil price as a predictor variable. The key econometric issue in the predictability literature is that if one uses ordinary least squares (OLS), which is typically the case, one ends up with bias inference on the null hypothesis of no predictability. This is because the OLS estimator performs poorly when subjected to a persistent and an endogenous predictor. It will be shown later that for many countries in our sample, the real oil price is not only persistent, it is also endogenous. These are not the only source of biasness though. In a recent study, Westerlund and Narayan (WN, 2012) show that the OLS estimator performs even more poorly if data are characterised by heteroskedasticity. Therefore, this paper makes an econometric contribution by modelling these salient features of the data by applying a recently developed generalised least squares (GLS) estimator proposed by WN (2012). The main feature of the GLS estimator is that it accounts for not only heteroskedasticity in the model but also the persistent and endogenous behaviour of the predictor variable. The countries studied are Bangladesh, Cambodia, China, Hong Kong, India, Indonesia, Korea, Malaysia, Myanmar, Japan, Singapore, Thailand, the Philippines, and Vietnam. These countries are especially interesting for two reasons. First, they add variety to the analysis because some of these nations are oil importers while others are oil exporters (see Table 1). They also tend to follow a mixture of exchange rate regimes, including free float, managed float, crawling peg, and Currency Board Arrangement (see Table 2). Therefore, we have at hand a set of Asian countries which are heterogeneous both in terms of the oil consumption/production, suggesting that oil price should affect them differently, and in terms of exchange rate regimes, suggesting that exchange rate responsiveness to oil price (shocks) may well be different. Table 1. Oil supply, petroleum consumption and trade of refined petroleum of 14 Asian economies. Total oil supplya,b Total petroleum consumptiona,c Exports less imports of refined petroleuma,d 90–09 90–99 00–09 90–09 90–99 00–09 90–09 90–99 00–09 Bangladesh 4 1 6 67 48 86 −40 −24 −57 Cambodia 0 0 0 18 8 27 −18 −8 −27 China 3382 3049 3715 4890 3294 6487 −356 −254 −458 Hong Kong 0 0 0 238 176 299 −237 −173 −300 India 760 698 822 2031 1525 2538 5 −254 264 Indonesia 1405 1597 1213 1002 807 1197 −29 110 −169 Japan 6 −12 25 5363 5552 5174 −1031 −1108 −954 Korea 108 90 126 1971 1773 2169 65 −61 192 Myanmar NA NA 17.56 30.05 21.27 38.83 NA NA NA Malaysia 737 712 763 444 377 511 −19 −69 30 Philippines 10 2 18 320 311 329 −56 −27 −85 Singapore 6 0 11 667 517 818 305 502 108 Thailand 194 97 291 741 615 867 21 −85 127 Vietnam 254 157 351 166 97 236 −160 −95 −225 a These are averages for thousand barrels per day. b Total oil supply includes the production of crude oil, natural gas plant liquids, and other liquids, and refinery processing gain. c Total petroleum consumption includes internal consumption, refinery fuel and loss, and bunkering. Also included, where available, is direct combustion of crude oil. d Refined petroleum products include but are not limited to gasoline, kerosene, distillates, liquefied petroleum gas, asphalt, lubricating oils, diesel fuels, and residual fuels. Sources: US Energy Information Administration (EIA). Data for Myanmar is extracted from www.indexmundi.com/energy. Table options Table 2. Exchange rate regimes and government regulation of oil price. Full sample period Exchange rate regimesa Presence of price subsidies on petrol, diesel and/or keroseneb Bangladesh 1999:05–2011:11 1983–1999: Pegged to a basket of major trading partners’ currencies. The US dollar was the intervention currency Yes. Prices are adjusted periodically to bring them closer to the world market price, but subsidies have remained substantial 2000–2003: An adjustable pegged system From 31 May 2003: Managed floating exchange rate system Cambodia 1997:06–2012:03 1992: Managed floating exchange rate system No China 1990:02–2012:03 Early 1990s: Rates adjusted on the basis of certain indicators such as performance of foreign exchange markets, export production costs and relative prices No 1994 onwards: managed floating exchange rate system Hong Kong 1990:02–2012:03 Since 1983: Currency Board Arrangement to fix Hong Kong dollar to US Dollar. As of April 1999, the fixed rate was HK and 7.80 per US$1 No India 1990:02–2012:03 1990s: A managed float exchange regime with the effective rate of rupee linked to a basket of currencies of India's major trading partners Yes. Deregulation of some petroleum products since April 2002 March 1992: Dual (official and market based rate) – the Liberalised Exchange rate Management System (LEMS) March 1993 The unified market based exchange rate system replaced LEMS Indonesia 1991:12–2012:03 Up to July 1997: Managed floating exchange rate system Yes. To consumers only From August 1997: Free floating exchange rate system Japan 1990:02–2012:03 Free floating exchange rate system Yes Korea 1990:02–2012:03 1980: Managed floating exchange rate system with a trading band No Since 1997: Free floating exchange rate system Myanmar 1994:09–2012:03 2012: Reference exchange rate under managed floating exchange rate system. – Before that the kyat was pegged to IMF Special Drawing Rights (SDRs). Malaysia 1990:02–2012:03 Since 1998: Managed floating exchange rate system but has been fixed to the US dollar Yes. Petrol rebate for vehicles, diesel subsidies for transport operators, fisheries Philippines 1990:02–2012:03 Since 1984: Floating exchange rate system No Singapore 1990:02–2012:03 1985: Floating exchange rate but called monitoring band. The Singapore dollar is monitored against an undisclosed basket of currencies of major trading partners and competitors No Thailand 1991:12–2012:03 1978: Pegged to a basket of currencies but allowed to float within a band Yes 1997: Managed float exchange rate Vietnam 1999:04–2012:03 Until 1989: Multiple rates were merged to make way for the convertible currency rate Yes. To state owned companies only 1991: Replaced by Auction fixing rate 1999: Only one dong fixed to the US dollar; commercial banks set their own rate within a band – crawling peg a Source: Exchange rate: IMF, 2010. Annual Report on Exchange Rate Arrangement and Exchange Restriction. b Source: Oil regulations: Jha, S., Quicing, P., Camingue, S. F. 2009. Prevalence of oil taxes and subsidies in developing Asia’ in macroeconomic uncertainties, oil subsidies and fiscal sustainability in Asia. ADB Economic Paper Series; and IMF, 2008. Fuel and food price subsidies: issues and reform options. Table options Second, an interesting feature of half of the countries in our sample is that they employ oil price subsidy on petroleum products (petrol, diesel and kerosene). This subsidy has come under scrutiny since 2003 when oil prices began to rise. For several Asian countries, the subsidy enables consumers and/or producers to purchase petroleum products at below world market price. While the key rationale for this policy is to make energy accessible to the poor, there are some countries, such as Vietnam, where the policy is targeted to assist producers. However, in many of these countries a higher oil price has meant this policy has weighed heavily on government budgets. Some countries, in particular Indonesia, Bangladesh, and India have considered reducing their subsidies, but this remains a difficult policy decision given strong public support for the policy. Some analysts link oil price subsidies with high oil prices. They believe this policy has created distortions in demand and supply of petroleum products and, hence, has contributed to a high oil price in the world market. While there is no empirical evidence to substantiate this claim, at least theoretically this is possible. A price subsidy does artificially keep demand for imported oil high. And, if a number of large countries implement this policy, then demand for oil would be substantially high enough to put upward pressure on international oil prices. A price subsidy inflated oil price may also contribute to (pronounced) depreciation of the local currency of oil importing countries against the US dollar.1 Since this study tests the ability of oil prices to forecast exchange rate returns for countries that have adopted this policy, the effects of an oil price subsidy is indirectly accounted for in our analysis. The rest of the paper is organised as follows. Section 2 provides an overview on demand and supply of oil, the exchange rate regimes, and the incidence of oil price subsidies in Asia. Section 3 explains the dataset. Section 4 discusses the predictive model. Sections 5 and 6 present the in-sample and out-of-sample forecasting results, respectively. The final section provides some concluding remarks.
نتیجه گیری انگلیسی
This paper examines whether oil prices can forecast exchange rate returns for a group of Asian crude oil importing and exporting countries faced with different exchange rate regimes. Some of these countries also have an oil price subsidy. This paper contributes to our understanding of the predictive role of oil prices in exchange rate returns and also demonstrates the use of the recently developed GLS test proposed by WN (2012) for testing the null hypothesis of no predictability. The main benefit of the GLS test is that it accounts for heteroskedasticity in the model, as well as predictor persistency and endogeneity. An out-of-sample forecasting evaluation, comparing the GLS estimator with the Lewellen estimator reveals that both estimators perform equally well in the in-sample analysis, although the out-of-sample analysis reveals that there are some countries that prefer one estimator over the other. The other key results emerging from the in-sample forecasting analysis is that oil price is an important predictor of exchange rates of Bangladesh, Cambodia, Hong Kong, and Vietnam vis-à-vis the US dollar. For Vietnam, we find a positive predictive slope coefficient, suggesting that high oil price today predicts future depreciations. Since Vietnam is a net oil importer, this result is consistent with theory. The other three nations are found to have negative predictive slope coefficients despite being net importers. Furthermore, the exchange rate–oil price relationship for Bangladesh and Vietnam is characterised by a policy of oil price subsidy. Given our research objective and econometric framework we are unable to directly test this policy. This is something we leave for future studies. Nonetheless, we studied several Asian countries (Bangladesh, India, Indonesia, Japan, Malaysia, Thailand, and Vietnam) which use this policy, but for which there is limited evidence that oil price predicts exchange rate returns.