人民币升值对中国贸易及经济的影响分析.doc

The Impact of RMBs Appreciation on Chinas Trade* Supported by NSFC and Ministry of Commerce of PR China. The authors are very grateful to the anonymous referees and the editor-in-chief for their very valuable comments and suggestions.Guihuan Zheng1,2, Li Guo2, Xuemei Jiang2, Xun Zhang1,2 and Shouyang Wang1,2,# The corresponding author, Email: ; Fax: 86-10-62621304. 1Institute of Systems Science, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100080, China2School of Management, Graduate University of Chinese Academy of Sciences, Beijing 100080, ChinaAbstract: With the announcement of reforming the RMB exchange rate regime on July 21,2005, great attention has been paid to the impacts of RMBs appreciation on Chinas trade. Since China being in the transition from pegging to a managed floating regime, no existing approximate model and method can be utilized to investigate this question directly. In this paper, scenario analysis technique is used to give a study on this issue, coupled with the introduction of the substituted valuables: Japanese yen and Euro exchange rate. Our results show that RMBs appreciation would not bring severe effects on Chinas trade in 2005; however, the possible sustained reduction of export growth in 2006 should be paid more attention. It is also necessary and urgent to press forward with the reform of RMB exchange rate regime, and put up the related supporting policies to avoid such sudden fluctuation as the Japanese situation after “Plaza Accord”.JEL Classification: C53, F17, F31Key words: RMBs appreciation, Chinas trade, Scenario analysis1. IntroductionThe RMB exchange rate became a hot topic after the announcement of exchange rate regime reform on July 21, 2005.The new regime, a managed floating exchange rate regime based on market supply and demand with reference to a basket of currencies, allows RMB to rise by 2%, with a daily 0.3% trading band based on the price of the previous day. With its rapid development, Chinas economy has become an increasingly important element in the global economic development and integration. Against this backdrop, the impacts of RMBs appreciation on Chinas trade attract great attention.Before this reform of RMB, there has emerged a vast theoretical literature concentrated on the impacts of RMBs revaluation (Liu,2004; Zhang,2001). However, relatively little research has been undertaken on the empirical analysis of this issue. Cyn-Young Park (2005), an economist in Asian Development Bank, analyzed macroeconomic impacts of a “one-off” appreciation of the RMB against the dollar using the Oxford Economic Forecasting (OEF) model. This work may shed light on the dynamics between global imbalances and revaluation. The OEF model framework (Burridge et al.,1991) allows simulation analyses based on the global econometric structure, which will provide some quantitative results for the impacts of a revaluation on the concerned economies, such as the PRC, Japan, US, and other Asian countries. The stability of equilibrium state may be damaged when the regime shifts, which, however, cannot be studied in this model. Zhang (2005), a researcher in Chinese Academy of Social Sciences, presented a study on this issue by the elasticity analysis. Both of these results suggested that the impacts of a revaluation on Chinas economy might change in the proportion of an appreciations scale. In fact, in the short term, the reasonableness of this proportionate result should be doubted.The main body of existing literature, however, has so far mostly centered on quantitative discussion describing the relation between trade and exchange rate (Arize,1996; Bailey et al.,1987; Cheng et al.,2003; Chou,2000; Cushman,1983; Gotur,1985; Koray and Lastrapes,1989; Mckenzie,1999; Sauer and Bohara,2001; Swift,2004; Viaene and de Vries,1992; Wang and Dunne,2003). There has been surprisingly little empirical work that focus on the exchange rate regime shift (Zheng,2005), especially from pegging to the US dollar to a managed floating regime. No existing appropriate model or method can be found to support this kind of research directly. Therefore, some indirect methods should be experimented for studying this issue with the substitute variables introduced.In this paper, scenario analysis technique is used to give a sensitivity analysis of Chinas trade to the appreciation of RMB. The first type of scenarios is a hypothetic scenario being related with the current appreciation of RMB by 2%; while the second one that belongs to the historical scenario is based on the historical event - the Japanese yens variability after subscribing “Plaza Accord”.The remaining part of this paper is organized as follows. The theory and method of scenario analyses are introduced in Section 2. Models and their evaluation are outlined in Section 3. Forecasting results from scenarios analyses are presented in Section 4. Section 5 concludes the paper.2. Scenario analysisScenario analysis (Committee on the Global Financial System,2000) is a kind of stress testing which serves to estimate potential extreme losses of a portfolio value and give helpful suggestions to the decision makers in risk management of a company or a financial institution. Generally, scenario is a means to explore the future economic situation, and identify what might happen and how an organization can act or react upon future developments.2.1 MethodIn our work, scenario analysis serves to estimate the changes in trade if the scenarios were to occur. The essence of this methodology is the creation of user-defined scenarios, fed into a calculation engine to produce estimates of the impacts. Moreover, such analysis should be based on the evaluated model. Generally, the scenarios in stress testing usually fall into three main types:(1) The typical scenario is to use predefined or set-piece scenarios that have proven to be useful in practice, such as one x% change of exchange rate may result a y% fall in the stock index.(2) The hypothetic scenario is to report the worst-case results through some subjective assumption based on one special event.(3) The historical scenario is to stimulate the historical track for estimating the impacts based on one actual event.After setting the scenarios and establishing the valuation functions, the values of trade would be estimated through running the model for each scenario.2.2 Set scenariosIn our paper, two types of scenarios, the second one and the third one are introduced. They are based on the assumption that RMB would appreciate against the dollar and other economies could maintain the current exchange rate regimes.(1) The hypothetic scenarioAs we all know, RMB has appreciated 2% with a daily 0.3% trading band based on the price of the previous day. To illustrate the effects caused by this regime shift, one hypothetic scenario named scenariois introduced. In this scenario, it is assumed that the RMB/USD exchange rate would appreciate to a threshold value by 2% in August 2005, and maintain such value until the end of the next year. Since the effect of the related factors changes on macroeconomic structure is less than that on portfolios, it is reasonable to assume that the domestic and international markets keep stable and should not suffer a sudden change in the short run. This scenario analysis about one-off appreciation will be discussed in such assumed macroeconomic background.(2) The historical scenarioSeen from the previous history, many cases are worth being used as reference for the similar situation, such as Japan, Germany and Indonesia. Especially the case of Japan after subscribing “Plaza Accord” is suitable for providing a good scenario.On September 22nd 1985, finance ministers from the worlds five biggest economies - the United States, Japan, West Germany, France and the UK - announced the Plaza Accord at the eponymous New York hotel. Each country made specific promises on economic policy: the United States pledged to cut the federal deficit, Japan promised a looser monetary policy and a range of financial-sector reforms, and Germany proposed tax cuts. All countries agreed to intervene in currency markets as necessary to get the dollar down. By the end of 1987, the dollar had fallen by 54% against both the D-mark and the yen from its peak in February 1985. Japanese yen experienced a severe appreciation process from the level of 45.53 exchange rate index in September 1985 to the level of 53.13 in December 1985, increasing by 20% (See Fig. 1) (Data are taken from International Financial Statistics in IMF). After then, the trend of Japanese yens appreciation lasted about ten years.Figure 1: Japanese Exchange Rate Index after Plaze AccordAfter investigating the data about export, import, and exchange rate in Japan over the period of 1982-1990, we find that Japanese yens appreciation hasnt caused the changes of the long equilibrium relation between export and exchange rate. Therefore, it is feasible to set scenarios with an assumption of the stable economic structure. However, with the different international environment and historical conditions, the regimes shifting from pegging to US dollar may damage the equilibrium. Thus, two historical scenarios are set for providing enough reference:(1) Scenario : In this scenario, the RMB would start to appreciate against USD in August 2005, experiencing the same appreciation routine from September 2005 to December 2006 as the Japanese yen from December 1985 to February 1987. However, the state of economic equilibrium keeps stable.(2) Scenario : This scenario experiences the same appreciation track as scenario , but the state of economic equilibrium is assumed to change with the absolute value of elasticity coefficient increased by 10%.All of these scenarios might be unlikely to happen in the future, but they can describe the stress of RMBs appreciation that Chinas trade and economy could endure.3. Models and their evaluationThe RMBs appreciation should undoubtedly generate extensive and far-reaching implication for Chinas economy, especially for exports. This research should be based on establishing and testing econometric models. We establish the evaluated models for export and trading forms respectively. In the following empirical work, these models are used in the scenario analysis of Chinas trade. The sample data are collected over the period from Jan. 1995 to Jul. 2005. Data on export and import are taken from Chinas Customs Statistics. The exchange rate data are obtained from the University of British Columbias website (http:/fx.sauder.ubc.ca/data.html).3.1 Constructing the valuation function of exports3.1.1 Total exportsThe export function, constructed by Wang et al. (2003) for discussing the effects of Japanese yens depreciation on Chinas exports via the stress testing analysis, is used in our research. It is reasonable to assume that the foreign income, the domestic and foreign price level keep constant in a small economy. So the effects of exchange rate on Chinas exports will be analyzed in such assumed macroeconomic background. That is: (1)where represents the Chinas exports, is the exchange rate of RMB.It is well known that Japanese yen is the main currency in Asia, even in the world. Its change will affect the bilateral trade between China and Japan, as well as exports to Asia and America. And all of these will affect the Chinas exports finally. Moreover, the exchange rate of RMB to US dollar has been kept relatively stable since 1994, while Japanese yen was characterized by the floatability and flexibility. These evidences show that it is appropriate to choose Japanese yen as the substitute variable of exchange rate. Besides, the bilateral trade between China and Europe becomes increasingly important. Chinas exports to EU reached US$ 107.16 billion in 2004, accounting for 18.1% of Chinas total exports. It is obvious that euro is the choice substitute variable. Therefore, the export equation can be rewritten as (2)where represents the monthly data series on Chinas exports, , are the nominal exchange rates of RMB to Japanese yen and euro respectively. The variable of export growth rate , is defined aswhere is the natural logarithm of. Because the growth rate of Chinas exports is defined as the rate over the corresponding period of the previous year, the seasonal adjustment in the series of Chinas exports need not be considered. The growth rate of RMB/YEN exchange rate is defined as (3)where is the natural logarithm of the exchange rate of RMB/YEN data series, and is defined similarly. The results of unit root tests on , and are displayed in Appendix Table A.1. All of these series are stationary.However, the coefficient of the exchange rate variable is insignificant in this study, which indicates that the impact of the euro on the total exports cant be seen evidently. It requires further study in terms of the main export markets such as Asia, EU, and USA. It is found that has lagged impacts on Chinas exports, and the lag is from 5 to 10 months. The other finding is the autoregressive feature in Chinas exports. Based on these phenomena, an ADL (Auto-regressive Distributed Lag) model is applied, and the estimated regression is as follows (the value in parentheses is t-statistic for each coefficient): (4) (2.7778) (6.0607) (4.9982) (2.9985) (2.9657), From the above results, we may also find the coefficients of variables are significant at 1% level. Consequently, our hypothesis of small economy assumption is reasonable in short-run.3.1.2 Main export marketsGenerally, the impacts of RMBs appreciation on Chinas exports to different markets may differ obviously due to the various trade patterns, commodity structure, trade term and so on. The systematic model should be utilized to estimate for the export persistence and substitution. Furthermore, the VAR model for Chinas exports to main markets is established, and only the exchange rate is considered as the exogenous variable like the above export model.Since Chinas exports to Asia, EU and USA account for almost 89% of Chinas total exports, only these three markets are considered in our model. Because the exchange rate of RMB is pegged to US dollar with a fluctuation less than 0.04% before July 22nd, 2005, Chinas exports to USA are mainly influenced by the RMBs changes to other currencies rather than USD. Therefore, the exchange rates of RMB/YEN and RMB/EURO are used in the model.The VAR model can be described as: (5)where is the vector of monthly growth rates of Chinas exports to Asia, EU and USA, that is . is the monthly growth rate of Chinas exports to Asia measured in twelfth difference of the natural logarithm of Chinas exports to Asia, and are defined similarly. is the vector of monthly change rate of and , that is . and are the lags for each variable in the model， and are matrices of parameters, is the constant vector and is the vector of error terms.The dataset used in this model is limited to the period from Jan. 1999 to Jul. 2005. We did not use data before Jan. 1999 because the robustness and stability of this model was influenced by Eastern Asian financial crisis in 1997. The model VAR(2) could be used because of its smallest AIC and SC among the models with different lag lengths. Its results are displayed in Appendix Table A.2. Seen from it, the adjusted R-squared value for each equation is around 0.70. It represents strong explanatory power.3.2 Constructing the valuation function of trading formsIn the research of the impacts of RMBs appreciation on import, we established the ADL model with domestic demand and exchange rate. Industry added value was treated as the substitute variable to represent the domestic demand because of the absence of monthly GDP data. Since the coefficient of exchange rate variable is insignificant, we conclude that factors affecting Chinas total import are complicated and it might better to do further research in trading forms.Foreign trade mainly consists of processing trade and general trade, and the effects of RMBs appreciation on Chinas foreign trade vary with the trading forms. Therefore, it is necessary to construct the models with the four variables: general trade export, general trade import, processing trade export and processing trade import. As in 3.1.1, and are considered as the substitute variables of exchange rate. We assume that foreign income (reflecting foreign market demand) and trade policy keep constant. For the interaction between export and import through processing trade, it is reasonable to apply VAR model that can be described as: (6)where is the vector of monthly growth rates of different trading forms, that is , where , , , denote the monthly growth rate measured in twelfth difference of the natural logarithm of Chinas processing trade export, processing trade import, general trade export and general trade import respectively. While is defined as 3.1.1, and is the change rate of (industry added value) measured as the following: (7), and are the lags for each variable in the model， and are matrices of parameters, is coefficient vector, is constant vector and is the vector of error terms. The estimation results over the period from Jan. 1995 to Jul. 2005 (reported in Appendix Table A.3) indicate that the adjusted R-squared values are above 0.7, which suggesting good explanatory power.4. Forecasting results from sce