To what extent does real exchange rate
volatility of Euro/CNY influence the
imports of Euro area from China?
Master thesis
Xiaocui Long
Student number: S2070219
Email: longxiaocui@yahoo.com.cn
Msc International Financial Management Msc Business and Economics Faculty of Economics and Business Faculty of Social Sciences
University of Groningen Uppsala University
Abstract
A large number of studies have discussed the relationships between exchange rate volatility and the trade flows; however, there is no consensus on the result. Moreover, among the studies, few literatures focused on the Euro area and China. This paper investigates how the real exchange rate volatility of Euro/CNY influences the imports of Euro area from China. By using the more recent data from January 1999 to December 2012, the results show that there is an insignificant relationship between exchange rate volatility of Euro/CNY and the imports of Euro area from China.
JEL Classification: F14, F31
1. Introduction
After the collapse of the Bretton Woods system in 1970s, many countries approved to utilize the floating exchange rate system1. Since then, there have been numerous studies about the impact of exchange rate fluctuations on international trade flows. However, these studies did not arrive at a consensus. In previous theory studies, Hooper and Kohlhagen (1978), Clark (1973), Broll (1994) and Wolf (1995) argued that the increase in exchange rate volatility will lead to the reduction of international trade volumes. This is mainly because that higher exchange risk result in lower risk-adjusted expected revenue from exports, and therefore it will reduce the exporter’s incentives to trade. However, authors like Franke (1991), De Grauwe (1988), and Giovannini (1988) predicted the opposite result as they thought that by using hedging instruments, traders can lower the risks. Along with the previous theory studies, evidences from empirical studies turn out different results as well. Godwin and Benson (2009), Byrne et al. (2008), Siregar and Rajan (2004) found that exchange rate volatility has a significant and negative impact on trade. Agolli (2003), Doyle (2001), Arize (1998b), McKenzie and Brooks (1997) found the opposite result and provided supports for the positive effects. Alam and Ahmed (2010), Bahmani-Oskooee and Payesteh (1993) reported that there is no significant relationship between these two factors.
Among these studies, I found that most of them, for example Arize (1998a), Cushman (1986), Gotur (1985) Caporale and Doroodian (1994), are mainly focused on the countries like US, UK and South Africa, few paid their attention to the Euro area. Although authors like Anderton and Skudelny (2001) and Arize (1998b) drew their attention to the European countries, however, their studies included countries not only from Euro area but also from European Union. Consequently, there are no study centers on only the whole Euro area. Therefore, this paper tries to fill the gap and makes distinction between the European countries and the Euro area and focuses on the whole Euro area. As it is pointed out by the European Commission that the trading
1
between European countries and China has become more and more important and China has become the second trading partner and the biggest source of imports of EU.2 At the same time, the EU is also important for China. It is one of the largest export markets for China. In 2012, China’s exports to EU account for 16% of the total exports to the world.3 It is only 1% lower than the biggest export partner United States. Moreover, the exchange rate between Euro and Chinese Yuan has changed frequently these years, for instance, the exchange rate of Euro to Chinese Yuan (Euro/CNY) hit 10.85 in July of 2008 and fell to 8.19 in the middle of 2010; and after that period it moved around 8.5 to 9.5 and then dropped to 7.90 in 2012.4 Considering the reasons above, I will focus on the Euro area and China and try to analyze the relationship between exchange rate volatility of Euro/CNY and the import of Euro area from China.
Therefore, the main research question of this paper is:
To what extent does real exchange rate volatility of Euro/CNY influence the imports of Euro area from China?
In order to answer the research question mentioned above, this paper will use monthly time series data from January 1999 to December 2012.
Although this paper only focuses on discussing the relationship between exchange rate volatility and imports of Euro area from China and it may lead to partially correct results, this paper is still important and it different from previous studies in several ways.
Firstly, in contrast to some previous studies, this paper makes clear distinctions between European countries and the Euro area, and only focuses on the countries from the Euro area. There are 17 countries, which are Belgium, Germany, Ireland, Spain, France, Italy, Luxembourg, the Netherlands, Austria, Portugal, Finland, Greece, Slovenia, Cyprus, Malta, Slovakia and Estonia, included in this study. This means that
2
Facts and figures on EU-China Trade. European Commission. September 2012.
3 http://www.tradingeconomics.com/china/exports. 4
this paper not only takes the big countries’ effects into account but also takes the small counties’ influences into consideration. Therefore, it will provide the importers in Euro area with a more clear and general view on how the exchange rate volatility of Euro/CNY influences the import volumes of Euro area from China.
Secondly, this paper uses monthly data instead of quarterly data. This improves the result and the result will be more accurate as the sample size becomes larger. Furthermore, this paper also provide some implications for the policy makers as Arize (1998a) summarized that the study of relationship between exchange rate fluctuations and trade is very important for the design of exchange rate and trade policies. This is mainly because that the policy that aims to increase trade volumes would become unsuccessful if exchange rate volatility leads to a reduction in trade volumes. In other words, as policy makers can influence the exchange rate regimes and policies, how the exchange rate volatility influences the trade volumes is important for them to make decision. At the same time, the intended effect of a trade liberalization policy may be doomed by the volatile exchange rate and this may also lead to a balance-of-payment crisis (Arize, 1998a; Arize et al., 2000). Therefore, by including 14 years data, from year 1999 to year 2012, both the time period before the change of Chinese exchange rate regime and the period after the change, this paper will provide the policy makers with some information on how the current exchange rate regime influences the trade volumes.
Therefore, this paper accounts for the effects from previous information and uses the GARCH model as a measure of exchange rate volatility.
In addition, in order to avoid the spurious regression that is happened while the model includes both stationary and non-stationary series (De Vita and Abbott, 2004); this paper utilizes the recently developed Autoregressive Distribute Lag (ARDL) model to test the long-run relationship between the variables. This model is simpler compared with the Engel and Granger (1987) two-step residual-based model and the Johansen (1991, 1995) maximum likelihood reduced rank model, because it does not require pre-testing procedure to ensure that all the regressors are I (1). And it is applicable irrespective of whether the regressors are purely I(0), purely I(1) or mutually co-integrated while required.
The following paper is structured as follows. Firstly, previous studies about the relationship between exchange rate volatility and trade will be reviewed. Secondly, the model that is used in this paper and the data are provided. Then, the empirical results are presented. Lastly, the conclusion will be given.
2. Literature review and hypothesis
There have been numerous studies discussing the relationship between exchange rate volatility and international trade flows. And they have never reached an agreement. In this chapter, the theories will come first and then followed by the reviewing of empirical studies.
2.1 Theories
risk averse, an increase in exchange rate risk raises the expected marginal utility of export revenue and induces them to export more. Moreover, as the risk-averse individuals worry about the worst possible outcome, in order to avoid the possibility of a drastic decline in their revenues when risk increases, they may decide to trade more.
Meanwhile, some authors (Arize, 1997; McKenzie, 1999) also concluded the relationship between exchange rate volatility and international trade is ambiguous. It is claimed that the degree of risk aversion is important to determine the effect of exchange rate uncertainty on trade. If an exporter is highly risk-averse, the expected marginal utility of export revenue will raise along with an increase in the exchange rate volatility. It is because that as the volatility increases, exporters may prefer to produce more so as to avoid a decline in the export revenue. Therefore, there will be a positive relationship between exchange rate volatility and trade volumes. However, if the exporters show a low degree of risk aversion, they prefer to export less since higher exchange rate volatility reduces the expected marginal utility of export revenue. Therefore, the exchange rate volatility may have negative or positive effects on exports and the theory cannot determine the relation between foreign trade and the volatility of exchange rates.
that the increase in exchange rate volatility has opposite effects on imports and exports as importers and exporters are on opposite sides of the forward market.
2.2 Empirical studies
significant positive relationship between exchange rate volatility and trade volumes by using the ARCH model. Besides, Arize (1998b) provided new evidence on the long-run relationship between imports and exchange rate volatility among eight European countries from the period of February 1973 to January 1995 by using Johansen's (1991, 1995) co-integration approach and robust single-equation methods. He found that exchange rate volatility has a positive and significant impact on the import volumes for Greece and Sweden whereas for the other six countries, the result is negative.
determined. On the one hand, some authors hold the view that the exchange rate volatility has no significant effect on trade volumes, e.g. Alam and Ahmed (2010) and Bahmani-Oskooee and Payesteh (1993), on the other hand, some authors support for a significant positive relationship, for instance, Agolli (2003), Doyle (2001), Arize (1998b) and McKenzie and Brooks (1997) etc. What is more, there are studies found negative effects, for example Hooper and Kohlhagen (1978), Clark (1973), Broll (1994) and Wolf (1995) etc.
The conflicting situation can also be reflected from a meta-analysis study from Coric and Pugh (2008). They tried to analyze the results and ranged the results from strong negative to strong positive effects. They found that 33 studies concluded that exchange rate volatility exerts an adverse effect on trade volumes and 25 studies concluded that this is not the case, moreover, 6 of those studies concluded that exchange rate variability is trade-enhancing (Coric and Pugh, 2008).
To summarize, the discussions above suggest that the impact of exchange rate volatility on trade volumes is an empirical issue and theory alone cannot stand (Arize, 1998b). As Bourdon and Korinek (2011) stated “research results which find positive, negative or no effect of exchange rate volatility on the volume of international trade are based on varied underlying assumptions and only hold in certain cases”, therefore, the hypothesis in this paper is:
H0: The real exchange rate volatility of Euro/CNY exerts a significant impact on the
imports of Euro area from China.
H1: The real exchange rate volatility of Euro/CNY exerts an insignificant impact on
the imports of Euro area from China.
3. Model and methodology
3.1 Model
exchange rate environment is from Gotur (1985) and the model is: ,
where is the quantity index of total manufacturing imports delivered; is the real domestic activity level; is the price of domestically produced substitutes for imported manufactures in domestic currency; is the price of foreign-produced manufacturing goods faced by domestic consumers, in domestic currency; and is the exchange rate risk facing demanders of imported goods.
This model was applied by many studies, like Arize (1998a, 1998b) and Arize and Shwiff (1998) etc. and it have been examined by the empirical study of Kenen and Rodrick (1986) and they found a favorable result.5 Therefore, I follow the model provided by Gotur (1985) and add a dummy variable to control the effect of the change of exchange rate regime of China in 2005. Since this paper only aims to test how the exchange rate volatility affects the imports, only the exchange rate volatility is independent variable, other variables are considered as control variables. Moreover, all the variables are presented in terms of logarithms as it allows imports to react proportionally in accordance with the change of explanatory variables (Kenen and Rodrick, 1986). Consequently, the model to investigate how the exchange rate volatility of Euro/CNY affects the imports of Euro area from China is:
(1) where denotes the logarithm of real imports at time t of Euro area from China, represents logarithm of the national income of Euro area at time t, is a measure of relative import price of the Euro area at time t, which is the real effective exchange rate of Euro to Chinese Yuan and is the exchange rate volatility at time t in logarithm. is the dummy variable to control the impacts of the change of Chinese exchange rate regime. is the constant term; , , and are the coefficients that present the relationships between explained variable
5 For more detail about how the authors examined the model and what the results are, please read the paper Kenen,
P. B. and Rodrik, D. (1986). Measuring and Analyzing the Effects of Short-term Volatility in Real Exchange Rates.
and the explanatory variables. The is the disturbance term.
The imports demand equation (1) therefore assumes that firstly there is no time lag in the adjustment of actual imports and import demands. Secondly, other factors, such as hedging and other policy changes that affect the imports, are not taking into account.6 Consequently, from the model above, the real imports is influenced by the national income of Euro area, real exchange rate of Euro/CNY, the exchange rate volatility of Euro/CNY and the change of exchange rate regime of China in July 2005.
As abovementioned in Chapter 2, previous theoretical and empirical studies have not arrived at a consensus on how the exchange rate volatility influences the trade volumes. Therefore, the impact of exchange rate volatility on imports is an empirical issue and the sign of the relationship between the volatility and foreign trade cannot be determined.
Regarding the effect of relative income, standard demand theory indicates that there is a positive relationship between domestic income and the real import demand. Because if the real consumptions raise with the increase of real income, with an unchanged distribution of income, more foreign products will be purchased. Empirical studies also provided supports for the positive relation. Akpokodje and Omojimite (2009) and Bourdon and Korinek (2011) concluded that a rise in national income will lead to an increase of the purchasing power of national consumers and therefore leads to a positive effect. Arize (1998b) found that if an increase in income leads to an increase in real investment, the investment goods not domestically produced will be bought from abroad and this will lead to an increase in import volumes.
On the other hand, previous studies (e.g. Arize, 1998a, 1998b; Arize and Shwiff, 1998, Alam and Ahmed, 2010) provided evidence for a negative relationship between imports and relative price. According to Podivinsky, et al. (2004), due to a variation in the nominal exchange rate or a different rate of inflation between two countries, an
6 As there are 17 countries included in the Euro area, other factors, for example the hedging and policies that are
increase in relative price will decrease real imports. Akpokodje and Omojimite (2009) explained that a depreciation of real exchange rate will cause a higher cost of imports, therefore, will decrease the import trade volumes.
3.2 Methodology
There are four important notes regarding the equation (1) on discussing the relationship between exchange rate volatility and imports. Firstly, before applying the GARCH (1, 1) model to measure the exchange rate volatility,7 the Engle (1982) test for ARCH effect should be used to verify that whether the model is suitable for the data.8 The null hypothesis of the ARCH effect test is that all q lags of the squared residuals have coefficient values that are not significantly different from zero. If the value of the test statistic excesses the critical value from the distribution, the null hypothesis is rejected and there is an ARCH effect and the GARCH (1, 1) model is appropriate for the data.
Secondly, in order to test whether the change of exchange rate regime in China in July 2005 produces structural instability, I use the Chow breakpoint test over the period of January 1999 to December 2012. The potential breakpoint is the July 2005, in which China experienced the change of exchange rate regime which may lead to structural changes. The test statistics is the F-statistic, which is based on the comparison of the restricted and unrestricted sum of squared residuals.9 If the probability of the result is smaller than the significant level 5%, we can conclude the change of exchange rate regime of China in July 2005 leads to structural changes and should be included in the model, otherwise, it should be exclude from the model.
Thirdly, while the ARDL bound test can be applied irrespective the regressors are purely I(0), purely I(1) or mutually co-integrated, it is still necessary to ensure beforehand that the levels of all the regressors are not higher than I(1). In order to
7
Please refer to section 4.2 Data collection for more detail.
8 For more information about model and how to test the ARCH effect, please read: Engle, R. (1982).
Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of the United Kingdom Inflation.
Econometrica, 50, pp. 377–403..
9
investigate the properties of the time series, the Augmented Dickey-Fuller (ADF) test is employed to determine the levels of the variables. The result is usually denoted by Xt ~ I(d). There are three components can be included while applying the ADF test,
namely intercept, trend and intercept and none.10 If the value of the calculated ADF-statistics is smaller than the value of the critical value from Fuller table, then the series is said to be stationary, otherwise it is non-stationary.
Lastly, after checking the stationary of the time series, I will test the co-integration between the variables. As Brooks (2008) stated that for stationary and unrelated data, if one is regressed on the other, the t-ratio and R2 would be quite low. However, while applying the standard regression techniques to non-stationary and unrelated data, the end result could be a regression that ‘looks’ good but valueless. Granger and Newbold (1974) also pointed out that an unsatisfied stationary assumption using the OLS method may lead to and unreliable statistical results. This is called as spurious regression.
In previous studies, the most commonly used approaches to test for co-integration are the Engle and Granger (1987) two-step residual-based procedure and the Johansen (1988, 1991) maximum likelihood reduced-rank approach. These two tests both require a certain degree of pre-testing to ascertain that all the explanatory variables are integrated of order one. It is necessary because in the presence of a mixture of I(0) and I(1) regressors, standard statistical inference based on conventional co-integration tests is no longer valid. Harris (1995) found that the trace and maximum eigenvalue tests from the Johansen procedure will become difficult to interpret if I(0) regressors presented in the model, because these variables are likely to generate spurious co-integrating relations with other variables. Rahbek and Mosconi (1999) also presented how stationary regressors in a Johansen-type framework generate nuisance parameters in the asymptotic distribution of the trace statistic.
Moreover, as Kremers et al. (1992) and Mah (2000) noted that the result of Johansen
10 For more information on the ADF model, please read the paper: Dickey, D. A. and Fuller, A. (1979).
Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of the American Statistical
(1988) method is not reliable while study has small sample size. Therefore, the reasons that I choose the ARDL model to test for co-integration are: firstly, this model can be applied for co-integration analysis irrespective of whether the regressors are purely I(0), purely I(1), or mutually co-integrated (Pesaran, et al., 2001). Secondly, it is unnecessary that the order of integration of the underlying regressors be ascertained prior to the testing (Pesaran, et al., 2001). Moreover, as Pattichis (1999), Mah (2000) and Tang and Nair (2002) noted that this model can be also used for small sample study. Lastly, the bound testing approach still can be used even when the explanatory variables are endogenous (Alam and Quazi, 2003 and Halim, et al., 2009).
To apply the bound testing approach, I start by modeling the equation (1) as a conditional ARDL-ECM: (2)
where and t are drift and trend components, and , , and are long-run coefficient matrices for , , and . The D is the dummy variable that controls for the regime change of Chinese exchange rate in July of 2005. Moreover, the dynamic structure of the first difference of the explanatory variables is set to ensure an absence of serial correlation in the estimated residuals (De Vita and Abbott, 2004).
The null hypothesis of ‘no co-integration’ is tested by the F-statistic for the jointly significance of the coefficients of the lagged levels in the error correction model (2), which refers to that H0: = = = = 0. According to Pesaran et al. (2001),
critical value, the null hypothesis is rejected and this means that there is a long-run relationship exists. However, if the F-statistic falls below the lower critical value, the null hypothesis cannot be rejected and this refers to no co-integration between the examined variables. Nonetheless, if the F-statistic value lies between the lower bound and the upper bound, it refers to an inconclusive result and the knowledge of the order of integration is required before making conclusive inferences.
According to De Vita and Abbott (2004) and Pesaran, et al. (2001), if there is a long-run relationship between the examined variables, the equation (2) can be reduced to a conditional long-run model when ∆ln(RI) = ∆ln(NI) =∆ln(Rex) =∆ln(Vol) = 0: (3) where: = , , , , , and the error process follows an IID (0. ). These long-run coefficients are estimated by the ARDL approach to co-integration. The first step is to run OLS on the conditional ECM of equation (2) and then using the lag selecting criteria to determine the optimal structure for the short-run dynamics. If co-integration is established, the error-correction model is appropriate for all non-stationary variables in the co-integration equation.
4. Sample and data
This section is started with a brief introduction of the developments of exchange rate regime as well as bilateral exchange rate movements and trade relationships of the Euro area and China, and then followed by how the sample data is obtained.
4.1 Developments of exchange rate regime and bilateral trade Exchange rate regime
to as Yuan, have changed several times in accordance with the development of Chinese economic structure. The Euro has been a floating currency since its inception in 1999, whereas for Yuan, China applied a strict policy of foreign exchange restrictions and controls before 1978. After 1970s, the system was gradually transformed to a dual exchange rate system, with a fixed and a market rate existing side-by-side. However, in 1994, China abandoned the dual system and changed to a formally managed floating system. This was in fact a strict peg against the US dollar since China fixed the external value of the Yuan against the US dollar at an exchange rate of 8.28 Yuan to 1 US dollar between 1995 and July 2005. In facing the pressures from many international economic institutions, for example, IMF and the European Union and the G7, which have repeatedly demanded a more flexible and quicker appreciation of the Yuan, in July 2005, China ended the peg and changed to a more flexible exchange rate regime. The change is described as a change of policy from a “crawling peg” to a “managed float plus” (Herrmann, 2010)11
.The Chinese government announced that the value of the Yuan would be set relative to a currency basket, which composed of the US dollar, Euro, Won, and Yen. And the Yuan was up evaluated by 2.1% against the US dollar and the bands of permissible daily movements increased to +/-0.3% (Herrmann, 2010). At the end of 2008, the value of Chinese Yuan to US dollar had been increased by roughly 20%, with the Yuan trading at 6.83 against 1 US dollar (Herrmann, 2010).
Bilateral exchange rate movements and trade flows
There is no direct exchange between Euro and Chinese Yuan. The exchange rate between Euro and Yuan is derived from the third currency US dollar. The first years after the Euro was established in January 1999 was characterized by a depreciation of the currency relative to the Chinese Yuan. For example, the average exchange rate of Euro to Yuan is around 0.118 in January 1999, and it changed to 0.156 Euro against 1 Yuan in June 2001. Since then, it tended to appreciate against the Chinese Yuan by
11
For more information about the developments of exchange rate regimes China, please read: Herrmann, C. (2010). Don Yuan: China’s “Selfish” Exchange Rate Policy and International Economic Law, (Eds) J. P. Terhechte.
European Yearbook of International Economic Law, pp. 31-51. Goldstein, M. (2004). Adjusting China’s Exchange
around 21% and reached to the point of EUR 0.089 per Yuan in October 2009. After that period, the Euro to Yuan kept stable and started to fluctuate around 0.101 to 0.110 during November 2009 to December 2012.12
The import of Euro area from China keeps a general increasing trend during the whole period from January 1999 to December 2012. The import from China rose from 75 billion in 2000 to 248 billion in 2008, and then followed by a downward trend and declined to 214 billion in 2009. From then on, it starts a new increasing trend and reached a new peak of 293 billion in 2011. As a result, the trade deficit with China had increased from 49 billion in 2000 to a peak of 170 billion in 2008, and fell to 156 billion in 2011. In the first six months of 2012, the imports from China account for 16% of the total imports of the Euro area and China became the second most important trading partner after the USA.13
4.2 Data collection
Previous studies use quarterly data most often; this paper uses the monthly data from Jan. of 1999 to Dec. of 2012 in order to increase the accuracy of the results. Because the result will be more accurate as the sample size becomes larger.
Dependent variable
Imports
The import of the Euro area is the dependent variable. Since the real import takes the inflation factor into account, whilst the nominal import does not, I use real imports rather than nominal import in order to achieve more precise result. In order to calculate the real imports of the Euro area, I first collect the data of nominal imports from the Statistical Data Warehouse of European Central Bank (ECB), and then deflate by the CPI of Euro area to get the real imports. The nominal import data that I collect from the ECB is in thousands of Euro.
12 The numbers are estimated according to the data of exchange rate of Euro against Yuan from the European
Statistics database.
13
Independent variable
Exchange rate volatility
As the exchange rate volatility cannot be obtained directly and previous studies hold different views on what kind of exchange rates should be used and how to measure it. One of the debates is the utilization of exchange rate. Researchers have not reached a common agreement on whether to use nominal exchange rate or real exchange rate. The nominal exchange rate was used most often in earlier studies, for example, Hooper and Kohlhagen (1978), and Thursby and Thursby (1987), both used nominal exchange rate in their study to measure volatility. However, some other researchers like Gotur (1985) and Tenreyro (2004) proposed that the real exchange rate is the most proper tool to measure volatility since it affects trade through price competitiveness. Nevertheless, there are also other researchers who provided evidence for the opinion that there is little difference to the result of using nominal or real measures to measure volatility (e.g. Qian and Varangis, 1994; McKenzie and Brooks, 1997). This paper follows the idea of the researchers Gotur (1985) and Tenreyro (2004), and uses real exchange rate instead of nominal exchange rate.
easily adjust, although not costlessly, short-term risk through forward market transactions, but it is much more difficult and expensive to hedge against long-term risk. Second, Jansen (1989) stated that this method is an unconditional measure of exchange rate volatility and it lacks a parametric model for the time varying variance of a time series. Third, as Arize (1997) pointed out that the calculation is ad hoc in nature because the choice of the order of the moving average process is arbitrary. Many authors used different order of the moving average, for example, Cushman (1983) used four; Koray and Lastrapes (1989) used twelve; and Chowdhury (1993) used eight. In addition, Pagan and Ullah (1988) also pointed out that this method could lead to an underestimation of the effect of exchange risks as it does not take some information set into account.
Other method that is used more often recently for estimating exchange rate volatility is the Generalised Autoregressive Conditional Heteroskedasticity (GARCH) model that was introduced by Bollerslev (1986). Kroner and Lastrapes (1993), Caporale and Doroodian (1994), Lee (1999), Doyle (2001) and Del Bo (2009) all applied the GARCH model in their studies. The GARCH model is the generalized ARCH model and allows the conditional variance to be dependent on previous own lags. The fitted variance is a function of a long-term average value and the past values of the information about volatility and the fitted variance (Brooks, 2008). By doing so, this model takes the relevant past information into account and allows volatility clustering, so that for example, large variances in the previous period would generate large variances in the future.
This paper chooses the GARCH model to estimate the exchange rate volatility instead of the moving sample standard deviation as the former one takes the long-term effects into account. According to Bollerslev et al. (1992) the GARCH (p, q) model can be viewed as a reduced form of a more complicated dynamic structure for the time varying conditional second order moments. The model is as follows:
where is the coefficient,
and denotes the real
exchange rate of Euro/CNY. The is the error term. The conditional variance is estimated by:
(5)
where is the conditional variance of real exchange rate and are the parameters to be estimated. is the squared residuals generated from equation (4), called ARCH term and it measures information about volatility in the previous period.
is the GARCH term representing the last period’s forecast variance. This
GARCH model states that the conditional variance of a time series is dependent upon the squared residuals of the process and has the advantage of including heteroskesticity into the estimation procedure of the conditional variance (Choudhry, 2005).
Control variables
Relative price
The relative price is the ratio of import price from China to the domestic price of Euro area, which is proxied by real exchange rate. I use real exchange rate because that it accounts for the inflation affects. The real exchange rate also cannot be generated immediately and directly, therefore, I first collect the nominal exchange rate from the European Statistics database, and then derive the real exchange rate of Euro/CNY (Rex) from the quoted nominal exchange rate of Euro/CNY (EX) through the following formulas:
(6)
International Financial Statistics database.
National income
Since there are many countries included in the Euro area, the national income of the whole area is difficult to find, therefore this paper uses the industrial production index as a proxy of the national income and this may lead to partial correct result. The industrial production index also can be found from European Statistics database.
Dummy variable
In order to control the influences of exchange rate regime change of China in July 2005, I include the dummy variable. It is set to equal to 0 before July 2005 and equal to 1 afterwards.
The data description
As I have mentioned above, the data are monthly and ranging from January 1999 to December 2012. Table 3 shows the basic characteristics of all variables in terms of raw data (not in logarithm).
Table 3: Data description
RI NI Rex VOL Mean 103263.4 103.4982 0.112936 0.001012 Maximum 171196.7 116.6100 0.160946 0.004069 Minimum 29276.69 93.97000 0.089449 0.000336 Std. Dev. 43726.50 5.270960 0.018912 0.000515 Skewness -0.043955 0.799552 0.957008 3.410206 Kurtosis 1.527884 2.935665 2.640340 16.60782 Observations 166 166 166 166
Note: RI, NI, Rex and Vol refer to real imports, national income, real exchange rate and exchange rate volatility, respectively. The real import is in thousands of Euro and as there is no direct data of national income of the whole Euro area, I used industrial production index as a proxy.
decline between 2008 and 2009, the real import of Euro area represents an increasing trend in general in the testing period. However, there is no clear trend of the national income, real exchange rate and the volatility (see Chart 2, 3 and 4 in Appendix); they waves among the sample period and there are obvious fluctuations during the year 2008 to 2010. This is probably due to the world wide spread financial crisis. In addition, the data of the volatility significantly deviate from the normal distribution as it has high skewness and kurtosis figures.
5. Results and analysis
5.1 Results of AHCH effects and the measurement of the exchange rate volatility
The result of the ARCH effects test is in Table 1. As we can see from the table that the computed value from the ARCH-LM test is 109.33, and the probability is 0.00, which is lower than 0.05, therefore, there exists an ARCH effect and we can conclude that the GARCH model is appropriate for the data.
Table 1: Result from ARCH-LM test
F-statistic 109.3297 Prob. F(12,143) 0.0000
Obs*R-squared 140.6676 Prob. Chi-Square(12) 0.0000
The result of GARCH (1, 1)14 model is shown in Table 2. As we can see from the Table 2 that the sum of approaches to 1. According to Engle and Bollerslev (1986), if + =1 in a GARCH (1, 1) model, there is persistence of a forecast of the conditional variance and infinite variance for the unconditional distribution of . In other words, the current shock indefinitely influences the future variance. As it is shown in Table 2, the sum of and is 0.84, which is very close to 1, this refers that the persistence of shocks to the exchange rate volatility is
14
great and the influence will decay very slowly. Table 2: Results from GARCH (1, 1) model
(0.501526) (-0.377057) (1.202698) (1.627136) (3.809251)
Log likelihood = 343.5468, Durbin-Watson statistic =2.116966, + =0.841029 Notes: z-statistics in the parentheses. *Significant at the 5% level.
5.2 Chow breakpoint test
The result of F-statistic from the Chow breakpoint test is 42.03 and the probability is 0.000, which is smaller than 5%. This refers that the null hypothesis of no structural breaks at the specified breakpoint is rejected. In other words, the change of Chinese exchange rate regime leads to economic change in China. Consequently, the dummy variable should be included to control the impacts from the change of exchange rate regime of China.
5.3 The unit root tests
As is shown in the Chart 1, 2, 3 and 4, all the variables except the chart for real imports, which presents a clear increasing trend in the whole period, fluctuate up and down during the sample period. Therefore, I include both trend and intercept components for the real imports while applying the ADF test, whereas for others only the intercept component is included. Table 4 represents the ADF tests results.
The Table 4 reports the ADF tests for a unit root in the logarithm of the real imports, the logarithm of national income, the logarithm of relative price and the logarithm of exchange rate volatility. It indicates that the variables ln(RIt), ln(NIt) and ln(Rext) are
integrated of order one, whereas ln(Volt) is integrated of order zero. This result justify
relationships between variables and confirms that the ARDL approach is the most appropriate and useful co-integration test in the context of this paper.
Table 4: Results of ADF unit root tests
Variables Integrated level ADF statistics 1% Critical value
Ln(RIt) I(1) -17.38137 -4.014288
Ln(NIt) I(1) -4.633377 -3.470427
Ln(Rext) I(1) -13.07769 -3.469933
Ln(Volt) I(0) -3.944898 -3.470179
Note: The lag order for the series is determined by Schwarz information criterion (SBC). While applying the ADF test, for all explanatory variables, ln(NIt), ln(Rext) and ln(Volt), the intercept
component is included, whereas for the explained variable ln(RIt), both the intercept and trend component are included.
5.4 Co-integration analysis
To test for co-integration, the bound test equation (2) is applied. Drift and trend components as well as dummy variables are included in the model.
Table 5: F-Statistics for Testing the Existence of Long-Run Relationship
Computed F-statistic Critical Value Bound Decision rule
8.3658 3.47 ~ 4.45 Reject null
Note: The critical value is taken from Pesaran, et al. (2001), unrestricted intercept and unrestricted trend with 3 regressors. Critical bound is based on 10 percent significant level.
Firstly, I choose to include 12 lags as I use monthly data in this paper, and then select the ‘optimal’ dynamic structure based on the Schwarz Bayesian Criteria (SBC). The optimal structure is ARDL (2, 1, 3, 0) based on the SBC and result of F-statistic is shown in Table 5.
long-run relationship exists and the exchange rate volatility, national income, real exchange rate and dummy variable have a long-term impact on the real imports. Accordingly, this result allows us to proceed to the conditional long-run model of equation (3) and the ARDL long-run coefficients for the model are presented in Table 6 below.
Table 6: Results of Estimated Long-Run Coefficients using the ARDL Approach
Regressor Coefficient Standard Error T-ratio
Ln(NI) 1.1250 0.69707 1.6139 Ln(Rex) -0.68923 0.25677 -2.6843*** Ln(Vol) -0.05814 0.088850 -0.65744 Inpt 3.8237 2.5667 1.4898 T 0.0058426 0.0012302 4.7494*** D 0.096781 0.085284 1.1348
Note: Inpt, T and D refer to drift component, trend component and dummy variables, respectively. *significant at 10% level **significant at 5% level ***significant at 1% level
increase in real exchange rate, the real import volume of Euro area will decrease by 0.69% in the long-run. It complies with the previous results of Akpokodje and Omojimite (2009) and Podivinsky, et al. (2004) that with the depreciation of real exchange rate, as the cost increases, the imports will decrease.
Table 7: Results of Error Correction Models
Regressor Coefficient Standard Error T-ratio
dln(RI1) -0.23650 0.077950 -3.0340*** dln(NI) 1.1279 0.38051 2.9642*** dln(Rex) 0.10501 0.10855 0.96738 dln(Rex1) 0.47679 0.11031 4.3224*** dln(Rex2) 0.42905 0.11556 3.7127*** dln(Vol) -0.011409 0.015559 -0.73326 dInpt 0.74681 0.47757 1.5638 dT 0.018902 0.4999E-3 2.2828** dD 0.023647 0.017922 1.0547 ECMt-1 -0.19531 0.058294 -3.3504***
ECM = ln(RI) - 1.1250*ln(NI) + 0.68923*ln(Rex) + 0.58414*ln(Vol) - 3.8237-0.0058426*T - 0.096781*D Note: dLn(RI) = Ln(RI)-LnRI(-1); dLn(RI1)= Ln(RI)(-1)-Ln(RI)(-2);
dLn(Vol) = Ln(Vol)-Ln(Vol)(-1); dLn(Rex) = Ln(Rex)-Ln(Rex)(-1)
dLn(Rex1) = Ln(Rex)(-1)-Ln(Rex)(-2); dLn(Rex2) = Ln(Rex)(-2)-Ln(Rex)(-3)
dLn(NI)= Ln(NI)-Ln(NI)(-1); dInpt = Inpt-Inpt(-1) dT = T-T(-1); dD = D-D (-1)
*significant at 10% level **significant at 5% level ***significant at 1% level
the ARDL co-integration test is -0.195, with a probability of 0.00, which is smaller than 5%. This implies that about 20% of the disequilibrium of the previous month’s shock adjusts back to equilibrium in the current month. Besides, the negative significant coefficient is not only an indication of adjustment toward equilibrium but also an alternative way to support the result of co-integration bound test that there is long-run relationship between the variables (Bahmani-Oskooee and Ardalani, 2006). From the Table 7, we can also find that the exchange rate volatility has a negative influence on imports in the short-run. The result implies that with 1% increase of exchange rate volatility, the import volumes will decrease by 0.58%. Meantime, the result of real exchange rate also turns out a negative impact on the real imports, with 1% depreciation of Euro against Yuan, the real imports will decrease by 0.69%. However, the national income depicts a positive effect: the 1% increase of national income will lead to 1.13% increase of the imports.
Sensitivity test
In order to discuss the sensitivity of the results, I test the results in two ways. Firstly, I divide the previously used data into two groups to verify whether the volatility produces different results, namely the data with high exchange rate volatility and the data with low exchange rate volatility. The original data is divided based on the mean value of the volatility, which is 0.001012. Consequently, there are 51 observations in the high volatility group and 117 observations in the low volatility group. Secondly, I separate the whole data into another two groups: the data before the change of exchange rate regime of China in July 2005 and the data afterwards. Then I repeat the ARDL process. The results are shown in Table 8.
relationship is inconclusive as the computed F-statistic falls between the critical bound.
Table 8: F-Statistics for Testing the Existence of Long-Run Relationship of Subgroups Data Group Computed F-statistic Critical Value Bound Decision
High volatility 1.5485 3.47 ~ 4.45 Accept null
Low volatility 3.5735 3.47 ~ 4.45 Inconclusive
Before Change 3.4434 3.47 ~ 4.45 Accept null
After Change 5.0447 3.47 ~ 4.45 Reject null
Note: The critical value is taken from Pesaran, et al. (2001), unrestricted intercept and unrestricted trend with 3 regressors. Critical bound is based on 10 percent significant level.
From the Table 8, we also find that the results of the group before and after the change of exchange rate regime of China in July of 2005 are different. For the group before the change, the F-statistic is 3.4434, which is smaller than the lower bound value 3.47, implies that the null hypothesis of no co-integration between the variables is accept. Nevertheless, the F-statistic is 5.0447 for the group after the change and it is higher than the upper bound value 4.45 and indicates a long-run relationship between the variables. Therefore, the results of the ARDL long-run coefficients for the group after the change are shown in Table 9.
Table 9: Results of Estimated Long-Run Coefficients using the ARDL Approach for Data Group after Change of Exchange Rate Regime of China in July 2005
Regressor Coefficient Standard Error T-ratio
Ln(NI) 0.90890 0.72848 1.2477
Ln(Rex) 1.0002 0.44119 2.2672**
Ln(Vol) -0.039477 0.088619 -0.44546
Inpt 9.5681 3.4115 2.8046***
T 0.0023913 0.0020520 1.1654
In Table 9, the statistics of exchange rate volatility and national income reach the same conclusions as the statistics shown in Table 6. However, the real exchange rate produces a significant positive effect on real imports after the change. It is different from previous result derived from Table 6. In other words, after the change of exchange rate regime in China, with 1% increase of real exchange rate of Euro/Yuan, the real imports will increase about 1% as well.
Table 10: Results of Error Correction Models for the Data Group after Exchange Rate Regime Change of China in July 2005
Regressor Coefficient Standard Error T-ratio
dln(NI) 1.2256 0.53645 2.2846** dln(Rex) 0.27386 0.12997 2.1070** dln(Vol) -0.010808 0.23335 -0.46318 dInpt 2.6197 1.0305 2.5420** dT 0.6547E-3 0.6611E-3 0.99035 ECMt-1 -0.27379 0.082617 -3.3140***
ECM = ln(RI) – 0.90890*ln(NI) – 1.0002*ln(Rex) + 0.039477*ln(Vol) – 9.5681-0.0023913*T
Note: dLn(RI) = Ln(RI)-LnRI(-1); dLn(Vol) = Ln(Vol)-Ln(Vol)(-1); dLn(Rex) = Ln(Rex)-Ln(Rex)(-1); dLn(NI)= Ln(NI)-Ln(NI)(-1); dInpt = Inpt-Inpt(-1); dT = T-T(-1);
*significant at 10% level **significant at 5% level ***significant at 1% level
short-run, the real exchange rate has a positive impact on real imports and with 1% depreciation of Euro against Yuan, the import volume will decrease by around 1%. In addition, there is a negative relationship between exchange rate volatility of Euro/CNY and real imports of Euro area from China, with 1% increase of exchange rate volatility, the imports will drop by 0.04%.
To summarize, while comparing the results for the whole data group in Table 6 and the results from the sensitivity test in Table 9, we find that in the long-run, no matter for the whole period from January 1999 to December 2012 or for the subperiod from August 2005 to December 2012, the exchange rate volatility of Euro/CNY depicts an insignificant impact on the real imports of Euro area from China. And it is the case for national income as well. However, the result for the real exchange rate is a bit different. Although the statistics from Table 6 and Table 9 all show significant results, the impacts are different. The real exchange rate has a negative impact on real imports for the whole period while the result turns out to be positive for the subperiod. While thinking about the short-term effects, both the results of national income and the volatility in Table 10 show the same sign as the results in Table 7. However, the real exchange rate shows a different result again. The result is negative for the whole period whereas it is positive for the subperiod. There is an interesting finding from the sensitivity test: with 1% depreciation of Euro against Yuan, the import volume will decrease by around 1% no matter in the long-run or in the short-run. Consequently, we can conclude that the change of the exchange rate regime of China do has some influences on the relationships between the variables, especially for the relationship between real exchange rate of Euro/CNY and real imports of Euro area from China.
6. Conclusion
makers to pay attention to these relations while making policy changes. What is more, it is also a big issue for managers to learn how the real exchange rate changes in the future and what the influences of current exchange rate policy are since these two factors influence the import demand directly and indirectly.
Acknowledgements
References
Akhtar, M. and Hilton, R. S. (1984). Effects of Exchange Rate Uncertainty on German and US Trade. Federal Reserve Bank New York Quarterly Review, pp. 7–16. Agolli, M. (2003). Exchange Rate Volatility Effect on Trade Variations. Albanian
Center for International Trade. Available at: http://pdc.ceu.hu/archive/00002085/.
Akpokodje, G. and Omojimite, B. U. (2009). The Effect of Exchange Rate Volatility on the Imports of ECOWAS Countries. The Social Science, 4(4), pp. 340-346.
Alam, M. I. and Quazi, R. M. (2003). Determinant of Capital Flight: An Econometric Case Study of Bangladesh. International Review of Applied Economics, 17, pp. 85-103.
Alam, S. and Ahmed, Q. M. (2010). Exchange Rate Volatility and Pakistan´s Import Demand: An Application of Autoregressive Distributed Lag Model. International
Research Journal of Finance and Economics, 48, pp. 7–22.
Anderton, R. and Skudeldy, F. (2001). Exchange Rate Volatility and Euro Area Imports. European Central Bank working paper series, 64, pp. 1-32.
Aristotelous, K. (2001). Exchange-rate Volatility, Exchange-rate Regime, and Trade Volume: Evidence from the UK-US Export Function (1889-1999). Economic Letters, 72, pp. 87-94.
Arize, A. C. and Shwiff, S. S. (1998). Does Exchange Rate Volatility Affect Import Flows in G-7 Countries? Evidence from Cointegration Models. Applied Economic, 30 (10), pp. 1269-1276.
Arize, A. C. (1997). Conditional Exchange Rate Volatility and the Volume of Foreign Trade: Evidence from Seven Industrialized Countries. Southern Economic Journal, 64, pp. 235-254.
Arize, A. C. (1998a). The Effects of Exchange Rate Volatility on U.S. Imports: An Empirical Investigation. International Economic Journal, 12(3), pp. 31-40.
Arize, A. C. (1998b). The Long-Run Relationship between Imports Flows and Real Exchange Rate Volatility: The Experience of Eight European Economies.
International Review of Economics and Finance, 7(4), pp. 417-435.
Asseery, A. and Peel, D. A. (1991). The Effects of Exchange Rate Volatility on Export: Some New Estimates. Economics Letters, 37(2), pp.173–177.
Bahmani-Oskooee, M. and Payesteh, S. (1993). Does Exchange Rate Volatility Deter Trade Volume of LDCs? Journal of Economic Development, 18 (2), pp. 189-205. Bahmani-Oskooee, M. and Ardalani Z. (2006), Exchange Rate Sensitivity of US Trade Flows: Evidence from Industry Data. Southern Economic Journal, 72, pp. 542-559.
Bailey, M., Tavlas, G. and Ulan, M. (1986). Exchange-rate Variability and Trade Performance: Evidence from the Big Seven Industrial Countries. Weltwirtschaftliches
Archiv, 122(3), pp. 466–477.
Bailey, M., Tavlas, G. and Ulan, M. (1987). The Impact of Exchange-rate Volatility on Export Growth: Some Theoretical Considerations and Empirical Results. Journal of
Policy Modeling, 9(1), pp. 225–243.
Bollerslev, T., Chou, R. and Kroner, K. (1992). ARCH Modeling in Finance. Journal
of Econometrics, 52, pp. 5–59.
Brada, J.C., and Mendez, J.A. (1988). Exchange Rate Risk, Exchange Rate Regime and the Volume of International Trade. KYKLOS, 41, pp. 263–280.
Broll, U., (1994). Foreign Production and Forward Markets. Australian Economic
Papers, 33, pp. 1–6.
Brooks, C. (2008). Introductory Econometrics for Finance. Second Edition.
Cambridge University Press, New York. pp. 392-394.
Byrne, J., Darby, J. and MacDonald, R. (2008). US Trade and Exchange Rate Volatility: A Real Sectoral Bilateral Analysis. Journal of Macroeconomics, 30 (1), pp. 238-259.
Caporale, T. and Doroodian, K. (1994). Exchange Rate Variability and the Flow of International Trade. Economics Letters, 46, pp.49-54.
Cho, G., Sheldon, I.M. and McCorriston S. (2002). Exchange Rate Uncertainty and Agriculture Trade. American Journal of Agriculture Economics, 84 (4), pp. 931-942. Choudhry, T. (2005). Exchange Rate Volatility and the United States Exports: Evidence from Canada and Japan. Journal of the Japanese and International
Economies, 19, pp.51-71.
Chowdhury, A. R. (1993). Does Exchange Rate Volatility Depress Trade Flows? Evidence from Error-correction Model. Review of Economic s and Statistics, 75, pp. 700 -706.
Clark, P. B. (1973). Uncertainty, Exchange Risk, and the Level of International Trade.
Western Economic Journal, 11(3), pp. 302-313.
Cushman, D. O. (1983). The Effects of Real Exchange Rate Risk on International Trade. Journal of International Economics, 15, pp. 45-63.
Cushman, D. O. (1986). Has Exchange Risk Depressed International Trade? The Impact of Third-Country Exchange Risk. Journal of International Money and
Coric, B. and Pugh G. (2008). The Effects of Exchange Rate Variability on International Trade: A Meta-regression Analysis. Applied Economics, pp. 1-14.
De Grauwe, P. (1988). Exchange Rate Variability and the Slowdown in Growth of International Trade. IMF Staff Papers, 35, pp. 63-84.
De Grauwe, P. and de Bellefroid, B. (1987). Long run Exchange Rate Variability and International Trade, (Eds) S. Arndt and J. D. Richardson. Real Financial Linkages
among Open Economies, The MIT Press, pp. 193–212.
Del Bo C. (2009). Foreign Direct Investment, Exchange Rate Volatility and Political Risk. European Trade Study Group annual conference, pp. 1-32.
De Vita, G. and Abbott, A. (2004). The Impact of Exchange Rate Volatility on UK Exports to EU Countries. Scottish Journal of Political Economy, 51(1), pp. 62-81. Doyle, E. (2001). Exchange Rate Volatility and Irish-UK Trade: 1979-1992. Applied
Economics, 33 (2), pp. 249-265.
Engle, R. (1982). Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of the United Kingdom Inflation. Econometrica, 50, pp. 377–403.
Engle, R. and Bollerslev, T. (1986). Modeling the Persistence of Conditional Variance.
Economic Review. 5, pp. 1–50.
Engle, R. F. and Granger, C. W. J. (1987). Cointegration and Error-correction Representation, Estimation, and Testing. Econometrica, 55, pp. 251-276.
Ethier, W. (1973). International Trade and the Forward Exchange Market. American
Economic Review, 63, pp. 494-503.
Franke, G., (1991). Exchange Rate Volatility and International Trading Strategy.
Journal of International Money and Finance, 10, pp. 292–307.
Frankel, J. A. and Wei, S. J. (1993). Trade Blocs and Currency Blocs. NBER Working
Giovannini, A., (1988). Exchange Rates and Traded Goods Prices. Journal of
International Economics, 24, pp. 45–68.
Godwin, A. and Benson, U. O. (2009). The Effect of Exchange Rate Volatility on the Imports of ECOWAS Countries. The Social Sciences, 4 (4), pp. 340–346.
Goldstein, M. (2004). Adjusting China’s Exchange Rate Policies. Institute for
International Economics Working Paper, No. 04-1, pp. 1-54.
Gotur, P. (1985). Effect of Exchange Rate Volatility on Trade: Some Further Evidence.
IMF Staff Papers, 32, pp. 475–521.
Granger, C.W. J. and Newbold, P. (1974). Spurious Regressions in Econometrics.
Journal of Econometrics, 2, pp. 111-120.
Halim, A., Ahmad, H., Daud, S. N. M. and Marzuki A. (2009). Sovereign Credit Ratings and Macroeconomic Variables: An Empirical Analysis on Dynamic Linkages in Malaysia. USIM Lecturer's Conference Papers, pp. 1-9.
Harris, R. (1995). Using Cointegration Analysis in Econometric Modelling. London:
Prentice Hall/Harvester Wheatsheaf.
Herrmann, C. (2010). Don Yuan: China’s “Selfish” Exchange Rate Policy and International Economic Law, (Eds) J. P. Terhechte. European Yearbook of
International Economic Law, pp. 31-51.
Hooper, P. and Kohlhagen, S. W. (1978). The Effect of Exchange Rate Uncertainty on the Price and Volume of International Trade. Journal of International Economics, 8, pp. 483–511.
Jansen, D. W. (1989). Does Inflation Uncertainty Affect Output Growth? Further Evidence. Federal Reserve Bank of St. Louis Review, pp. 43–54.
Johansen, S. (1988). Statistical Analysis of Cointegrating Vectors. Journal of
Johansen, S. (1991). Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregressive Models. Econometrica, 59, pp. 1551-1580.
Johansen, S. (1995). Likelihood-based Inference in Cointegrated Vector Autoregressive Models. Oxford University Press, Oxford.
Kenen, P. B. and Rodrik, D. (1986). Measuring and Analyzing the Effects of Short-term Volatility in Real Exchange Rates. Review of Economics and Statistics, 68, pp. 311-315.
Klein, M. (1990). Sectoral Effects of Exchange Rate Volatility on United States Exports. Journal of International Money and Finance, 9, pp. 299-308.
Koray, F. and Lastrapes, W. (1989). Real exchange rate volatility and US bilateral trade: a VAR approach. Review of Economics and Statistics, 71, pp. 708–712.
Kremers, J. J. M., Ericsson, N. L. and Dolado, J. (1992). The Power of Cointegration Tests. Journal of Econometrics, 52, pp. 389-402.
Kroner, K. and Lastrapes, W. (1993). The Impact of Exchange Rate Volatility on International Trade: Reduce from Estimates Using the GARCH-in-mean Model.
Journal of International Money and Finance, 12, pp. 298–318.
Lee, J. (1999). The Effects of Exchange Rate Volatility on Trade in Durables. Review
of International Economics. 7, pp. 189–201.
Mah, J. S. (2000). An Empirical Examination of the Disaggregated Import Demand of Korea-the Case of Information Technology Products. Journal of Asian Economics, 11, pp. 237 –244.
McKenzie, M. D. and Brooks, R. D. (1997). The Impact of Exchange Rate Volatility on German-US Trade Flows. Journal of International Financial Markets, Institutions
and Money, 7 (1), pp. 73-87.
McKenzie, M.D. (1999). The Impact of Exchange Rate Volatility on International Trade Flows. Journal of Economic Surveys, 13, pp. 71-106.
Pagan, A. and Ullah, A. (1988). The Econometric Analysis of Models with Risk Terms. Journal of Applied Econometrics, 3(2), pp. 87-105.
Pattichis, C. A. (1999). Price and Income Elasticities of Disaggregated Import Demand: Results from UECMs and Application. Journal of Applied Econometrics, 31, pp. 1061-1071.
Perée, E. and Steinherr, A. (1989). Exchange Rate Uncertainty and Foreign Trade.
European Economic Review, 33, pp. 1241–1264.
Pesaran, H. M., Shin Y. and Smith R.J. (2001). Bounds Testing approaches to the Analysis of Level Relationships. Journal of Applied Econometrics, 16, pp. 289-326. Podivinsky, J. M., Cheong, C. and Lu, M. (2004). The Effect of Exchange Rate Uncertainty on US Imports from the UK: Consistent OLS Estimation with Volatility Measured by an ARCH-type Model. Econometric Society 2004 Far Eastern Meetings, 957, pp. 1-16.
Qian, Y. and Varangis, P. (1994). Does Exchange Rate Volatility Hinder Export Growth? Empirical Economics, 19, pp. 371-396.
Rahbek, A. and Mosconi, R. (1999). Cointegration Rank Inference with Stationary Regressors in VAR Models. Econometrics Journal, 2, pp. 76–91.
Sercu, P. and Uppal, R. (2003). Exchange Rate Volatility and International Trade: A General-equilibrium Analysis. European Economic Review, 43, pp.429-441.
Siregar, R. and Rajan, R. S. (2004). Impact of Exchange Rate Volatility on Indonesia’s Trade Performance in the 1990s. Journal of the Japanese and International
Tang, T. C. and Nair, M. (2002). A Cointegration Analysis of Malaysian Import Demand Function: Reassessment from the Bound Test. Applied Economics Letter, 9, pp. 293-296.
Tenreyro, S. (2004). On the Trade Impact of Nominal Exchange Rate Volatility.
Federal Reserve Bank of Boston, Working paper, 03-2.
Thursby, J. G. and Thursby, M. C. (1987). Bilateral Trade Flows, the Linder Hypothesis, and Exchange Risk. Review of Economics and Statistics, 69, pp. 488–95. Viaene, J. M. and de Vries, C. G. (1992). International Trade and Exchange Rate Volatility. European Economic Review, 36, pp. 1311-1121.
Wolf, A. (1995). Import and Hedging Uncertainty in International Trade. Journal of
Appendix
15Chart 1: Real import
Note: the figures on horizontal axis are the period from 1999 to 2012 whereas the numbers in the vertical axis refers to the real import volumes in thousand Euros of Euro area from China.
Chart 2: Real exchange rate
Note: the figures on horizontal axis are the period from 1999 to 2012 whereas the numbers in the vertical axis refers to the real exchange rate of Euro to Chinese Yuan.
15
All the charts are derived by using the software Eviews 7. The information about the sources of the data and the calculation of the data, please read 4.2 sample and data.
Chart 3: National income
Note: IPI refers to industrial production index. It is a proxy of the national income of Euro area and collected from the European Statistics database. The figures on horizontal axis are the period from 1999 to 2012 whereas the numbers in the vertical axis refers to the industrial production index of Euro area.
Chart 4: Real exchange rate volatility
