The effect of China’s RMB

The effect of China’s RMB exchange rate movement on its agricultural export A case study of export to Japan Longjiang Chen College of Economics and Management, South China Agricultural University, Wushan, China Abstract Purpose – The paper aims to examine the relationship between changes and volatility of China’s Renminbi (RMB) exchange rates and its agricultural export. Design/methodology/approach – A GARCH(1, 1) model is specified to measure the exchange rate volatility and autoregressive distributed lag regression with structural break dummy variables is estimated based on the results of unit root test with structural break. Findings – The export supply model reveals that the net trade effect of RMB exchange rate movements relies on the comparison of exchange rate level change (appreciation or depreciation) effect and exchange rate risk effect. The empirical examination results, taking China’s agricultural exports to Japan as a case, show RMB depreciation against yen will promote export growth while appreciation hinder export, and exchange rate volatility positively stimulates agricultural exports to Japan. However, the effect of exchange rate volatility on the export is much smaller than that of exchange rate level, which leads to a negative net effect to the export. Originality/value – The constructed model and applied methodology contribute to a better understanding of the relationship between changes and volatility of China’s RMB exchange rates and its agricultural export. Keywords Exchange rates, Agriculture, Exports, China, Japan Paper type Research paper 1. Introduction Since the reform of Chinese Renminbi (RMB) exchange rate formation mechanism in July 2005, Chinese RMB has been in frequent fluctuations and appreciating gradually. In this context, close attention has been paid to the impact of the RMB exchange rate fluctuations on China’s export. Since each industry has its own characteristics, the exchange rate movements may have different effects on different industries. Therefore, as Klein (1990) pointed out, the impact of real exchange rate fluctuations on trade must be further investigated in the commodity level. In general, compared to manufactured goods, agricultural products have such special industry nature as lower initial cost of investment, the existence of long-term contracts, etc. It is generally acknowledged that the impact of exchange rate movements on agricultural trade is different from that on manufactured goods trade. Gue et al. (2002) has confirmed that, compared to other industries, the real exchange rate uncertainty has a more significant negative effect on agricultural trade. The current issue and full text archive of this journal is available at www.emeraldinsight.com/1756-137X.htm JEL classification – F14, Q17 CAER 3,1 26 China Agricultural Economic Review Vol. 3 No. 1, 2011 pp. 26-41 q Emerald Group Publishing Limited 1756-137X DOI 10.1108/17561371111103525 Then as for China, what is the effect of RMB appreciation and associated risks of volatility[1] on its agricultural export? A study on this issue has important academic and practical significance. On the one hand, there are still clear-cut theoretical and empirical conclusions on trade effects of exchange rate fluctuations among the economic community. Moreover, the vast majority studies have probed the trade effects of exchange rate fluctuations in developed countries and have neglected the study on those in developing countries like China. Developing countries have the characteristics different from the developed countries, such as the different exchange rate system, lack of a sound foreign exchange forward market, lack of effective financial derivatives, etc. which will impose important constraints on the behaviors of exporters. As for China, there are two important constraints: first, China runs particular foreign exchange management system; second, China’s agricultural exports have encountered serious technical barriers to trade (TBT) and sanitary and phytosanitary measures (SPS) from developed countries, owing to China’s low technical level and industrial standardization in agriculture. On the other hand, the current appreciation of RMB will bring certain negative impact on China’s agricultural economy and the welfare of farmers (Kong and Li, 2005), while farmers are still the majority of the population in China. Therefore, probing the effect of RMB exchange rate changes on the agricultural export has an extremely important practical significance. Based on the above understanding, this paper, taking China’s agricultural exports to Japan as a case, attempts to explain the impact of RMB exchange rate level changes and volatility on China’s agricultural exports. This paper is organized as follows. Section 2 is a brief review of the literature on the effect of exchange rate changes on agricultural trade. Section 3 constructs a model to analyze RMB exchange rate movements and agricultural exports facing two constraints including China’s particular exchange rate system and TBT/SPS in agricultural product trade. The model reveals that the net trade effect of exchange rate movements relies on the comparison of exchange rate level change (appreciation or depreciation) effect and exchange rate risk effect. Section 4 makes an empirical estimation. A GARCH(1, 1) model is specified to measure the risk of exchange rate volatility and autoregressive distributed lag (ADL) regression with structural break dummy variables is estimated based on the results of augmented Dickey-Fuller (ADF) unit root test, unit root test with structural break. Further analysis on the results shows that the dominant negative effect of continuing appreciation of RMB brings a negative net effect to the agricultural export to Japan. Section 5 summarizes the main conclusions and interprets policy implications. 2. Brief literature review The study on the impact of exchange rate changes on agricultural trade is launched first by Schuh (1974), which made the fundamental argument that the exchange rate was an omitted significant variable in economic analysis of the US farm sector. After his pioneering research, a considerable amount of researches evaluated quantitatively the impact of nominal and real exchange rate on agricultural trade. In recent years, the representative research results includes Susanti (2001), Gervais et al. (2004), Mathew et al. (2006), etc. Different empirical methods and exchange rate variables were used in these studies, but leading to the consensus: exchange rate fluctuations have significantly impeded agricultural trade flows, and compared to other sectors, such negative effect is more noticeable. Effect of China’s RMB exchange rate movement 27 Some empirical literatures compared the different effects of exchange rate changes on agricultural sector and other sectors. Both Gue et al. (2002) and Sheldon (2003) showed the different influences of exchange rate changes on machinery, chemicals, other manufacturing and agriculture. Compared to other sectors, the negative impact on agricultural trade is much greater than for total trade or for any other specific sector studied. Donald et al. (2005) found that the impacts of exchange rate volatility varied across sectors and were significantly negative on the agricultural trade. Some literatures only inspected the influences of exchange rate changes on agricultural sector trade, such as Jennifer (2006), Gu et al. (1994) and Song (2005). These studies reached similar conclusions, namely: overall, the exchange rate changes significantly negatively affected agricultural trade. Susanti (2001) and Mathew et al. (2006) investigated the effects of exchange rate changes on agricultural trade at both sector and product level. Susanti (2001) examined Indonesia’s total exports of agricultural products and five products export and revealed that all of them were significantly negatively affected by Indonesia’s exchange rate movements. Mathew et al. (2006) investigated the total agricultural exports and 12 agricultural products such as maize exports of the USA and drew similar conclusions of negative effects. Some researchers sought the evidence of negative effects of exchange rate movements on agricultural trade in the level of products. From the early investigation of Robert and Richard (1981) on American wheat, cotton and soybean to recent investigation of Jose et al. (2006) and Li and Li (2005) on soybean, the researchers investigated the impacts of exchange rate movements of a number of countries on their major agricultural trade flows including barley, wheat, pork, cotton, coffee, cocoa, living pig, corn, etc. These studies have generally supported the conclusion of the negative impacts of exchange rate movements on disaggregated agricultural trade. However, some studies (Anderson and Garcia, 1989; Abdulkudos, 2003) indicated that, for a certain agricultural product, the impacts of exchange rate movements vary across countries, and for a certain country, the impacts vary across agricultural products as well. In an era of floating exchange rate system, there are conflicting arguments in theory on if the volatility risk of exchange rate impedes international trade: some models find support for the negative hypothesis, yet other models have been derived to support the positive hypothesis. The results of empirical studies, which have focused on the exchange volatility, are less confusing but most of them support the negative effects. Sheldon (2003), Gervais et al. (2004), etc. indicated that the risk of exchange rate volatility significantly reduced agricultural trade flow. However, Jin et al. (2003) showed that whether the effect of exchange rate volatility is positive or negative is related to the measurement of volatility and the effect of third country as well[2]. A number of Chinese researchers have made investigations on the influences of RMB exchange rate changes on China’s agricultural trade with focus in two areas. One is to estimate J-curve effect and the exchange rate elasticity of import and export of agriculture-related products, including early studies of Cai (1994) and Gu et al. (1994) and recent studies of Song (2005) and Zhu et al. (2006). The other is to simulate the impact of RMB appreciation on China’s agricultural imports and exports within macro models, focusing the influences in different contexts of appreciation. Wei (2006) found the following results within a general equilibrium model: in a 5 per cent CAER 3,1 28 appreciation scenario, agricultural exports fall 0.9 per cent and imports increase by 1 per cent; in a 20 per cent appreciation scenario, agricultural export drops by 21 per cent and import increases by 8.1 per cent. Wang et al. (2005) showed that agricultural exports would be reduced by 580 million US dollars while imports increased by 2.8 billion dollars for 5 per cent appreciation of RMB. Although the above studies have achieved considerable success, there is still something that should be noticed. First, developing countries should receive more concern in future studies. There is still a big difference of the exchange rate system between developing and developed countries, and agricultural exports have played a more important role on developing countries’ economic development. Second, because of big differences between developed and developing countries, it seems that theory models need to be developed based on assumptions of developing countries. Third, close attention should be paid to possible structural break in data-generating process in order to prevent from spurious cointegration. 3. Model This section will construct a model on the basis of Kawai (1981), Fabiosa (2002), Barkoulas and Baum (2002) and Dekle and Jeong (2006). The impacts of both exchange rate level changes and the risk of exchange rate volatility will be incorporated into the model. 3.1 Monopolistic competition model Below we present the basic monopolistic competition model of the exporting firm. Let China be the export country and the foreign country be the import country. A representative household of the foreign country consumes differentiated goods that are arranged on the unit interval [0, 1]. Among the goods, goods z will be indexed for China’s agricultural product, where 0 , z , n, and goods z *[3]or foreign product, where n , z* , 1 (Figure 1). Assuming that domestic and foreign goods markets are segmented. The monopolistically competitive firm has the ability to engage in price discrimination by setting a domestic price for domestic sales that differs from the price it sets for exports (pricing-to-market). Based on the pricing-to-market model of Betts and Devereux (2000), we assume that the exporting firm sells hzt units of z output at time t in the home market and exports xszt to the foreign country. Total output of the exporting firm is: qzt ¼ hzt þ xszt ð1Þ The foreign representative household consuming differentiated goods seeks to maximize utility, following Dixit and Stiglitz (1977), the constant elasticity of substitution aggregate utility function at time t is: U ¼ Z 1 0 q ðu21Þ=u it di � �u=ðu21Þ ð2Þ where u . 1 is the elasticity of substitution between the varieties. Figure 1. Differentiated goods that are arranged on the unit interval Foreign product z*China’s agricultural product z 0 n 1 Effect of China’s RMB exchange rate movement 29 Let c*t ðzÞ be the foreign representative household’s consumption of the imported Chinese agricultural good z and c*t ðz *Þ be the consumption of the foreign good z *. Therefore, equation (2) can be rewritten as: U ¼ Z n 0 c*t ðzÞðu21Þ=udzþ Z 1 n c*t ðz *Þðu21Þ=udz * � �u=ðu21Þ ð3Þ The foreign money price index is: P*t ¼ Z n 0 p*t ðzÞ12udzþ Z 1 n p*t ðz *Þ12udz * � �1=12u ð4Þ where: p*t ðzÞ is the foreign currency price of good z and p*t ðz *Þ the foreign currency price of good z*. The foreign demand xdt for individual good z can be attained by maximizing U subject to total nominal expenditure I*t such that: xdt ¼ c*t ðzÞ ¼ p*t ðzÞ p*t " #2u I*t p*t ¼ p * t ðzÞ p*t " #2u R*t ð5Þ where R*t is the real income. 3.2 Export supply function The exporting firm uses domestic inputs and imported inputs to produce the export good z. Following Pick (1990), assuming a constant input-output ratio d for imported inputs kft , that is: kft ¼ d21qzt ð6Þ Thus, given linear cost function[4]: Cðwt;w*t ; qztÞ ¼ wtqzt þ w*t etd21qzt ð7Þ where wt is the price index for domestic inputs, w * t the price index for imported inputs and et the nominal exchange rate. Considering that TBT and SPS have severely hampered the growth of China’s agricultural export, TBT and SPS factors must be introduced into the model. Because of TBT and SPS, agricultural exporters have to pay additional costs in order to satisfy a new technology or other requirements, thus, the cost function is amended as follows: Cðwbt ;wt;w*t ; qztÞ ¼ wtqzt þ w*t etd21qzt þ wbt qzt ð8Þ where wbt is the shared unit additional cost in local currency due to TBT/SPS. Assume there is foreign exchange forward market to hedge for the firm facing exchange rate changes and the risks of unanticipated changes in exchange rate[5]. The exchange rate et is assumed to have a normal distribution[6], et~N ðe;s2eÞ. The firm’s net profit in RMB is defined as revenue less cost, that is: pt ¼ hztptðzÞ þ xsztp*t ðzÞet 2 ðwbt þ wt þ d21w*t etÞðhzt þ xsztÞ ð9Þ CAER 3,1 30 From equation (9), profit is also normally distributed with mean p and variance s2p given in equations (10) and (11): p ¼ hztptðzÞ þ xsztp*t ðzÞe2 ðwbt þ wt þ d21w*t eÞðhzt þ xsztÞ ð10Þ s2p ¼ ½xsztp*t ðzÞ2 d21w*t ðhzt þ xsztÞ�2s 2e ð11Þ The objective of the firm is to maximize the utility on the basis of profits. Assuming the firm is risk aversive, along with Fabiosa (2002), given a CARA utility function[7], the expected utility can be expressed in the form: EðvÞ ¼ 2e2l p20:5ls 2pð Þ ð12Þ It is a common result that the maximization of equation (12) can be equivalently expressed as: ðhzt ;xsztÞ Max {hztptðzÞ þ xsztp*t ðzÞe2 ðwbt þ wt þ w*t ed21Þðhzt þ xsztÞ 2 0:5l½xsztp*t ðzÞ2 d21w*t ðhzt þ xsztÞ�2s 2e } ð13Þ The solution to the first-order conditions gives: ptðzÞ2 p*t ðzÞeþ ls2ep*t ðzÞ½xsztp*t ðzÞ2 d21w*t ðhzt þ xsztÞ� ¼ 0 ð14Þ Using equation (14), comparative static analysis gives: ›xszt ›et ¼ 1 ls2e½p*t ðzÞ2 d21w*t � . 0 ð15Þ ›xszt ›s2e ¼ x s ztp * t ðzÞ2 d21w*t ðxszt þ hztÞ s2e½p*t ðzÞ2 d21w*t � ð16Þ That is, it is expected that an increase in the level of the exchange rate increases supply, while the impact of exchange rate volatility on export supply is not definitive, depending on the comparison of xsztp * t ðzÞ, the foreign exchange earnings of export, with d21w*t ðxszt þ hsztÞ, the total foreign exchange expenditure for imported inputs[8]. This indicates that the risk aversive firms respond differently to the increased exchange rate risk depending on the position where the firm is a net buyer or seller of foreign exchange. The firm facing increased risk of the exchange rate will reduce its export when the foreign exchange income of the firm is higher than the foreign exchange expenditure, while increase its export when less. But the firm will not respond to the increased risk of exchange rate when the foreign exchange earnings are exactly equal to the foreign exchange expenditure. Using the solution to the first-order condition in equations (14) and (5), the supply function for export good z is given by: xt ¼ xszt ¼ g þet; þ p*t ; þ R*t ; 2 wt; 2 w*t ; 2 wbt ;s 2 e � � ð17Þ where the signs of “ þ ” and “ 2 ” refer, respectively, to positive and negative elasticity, but for s 2e : Effect of China’s RMB exchange rate movement 31 ›xt ›s2e . 0 or ›xt ›s2e , 0 or ›xt ›s2e ¼ 0: That is, agricultural export volume is determined by four factors: (1) the nominal exchange rate et and associated exchange rate risk s 2 e ; (2) foreign real income R*t and the foreign market price level p * t ; (3) the price index for domestic inputs wt and the price index for imported inputs w*t ; and (4) wbt , TBT/SPS in agricultural trade. 4. Empirical estimation 4.1 Empirical specification To capture the dynamics, following Fang and Lai (2001), the empirical framework is specified as an ADL process combined with a GARCH(1, 1) model used for the measurement of exchange rate volatility risks: ln xt ¼ vþ Xa i¼1 liln xt2i þ Xb i¼0 ailn et2i þ Xc i¼0 bis 2 e;t2i þ Xd i¼0 diln p * t2i þ Xf i¼0 gilnR * t2i þ Xh i¼0 cilnwt2i þ Xj i¼0 wilnw * t2i þ Xk i¼0 hiw b t2i þ 1x;t ð18Þ et ¼ y 0 þ y 1et21 þ 1e;t ð19Þ s2e;t ¼ f0 þ f1 · 12t21 þ f2 ·s2e;t21 ð20Þ Equations (18)-(20) constitute two-step estimation with s2e;t generated by equations (19) and (20) and then used in equation (18) to estimate its effect on export. As Fang and Miller (2004) pointed out, the statistical significance and sign of the estimated ai and bi coefficients in equation (18) provide a simple and straightforward test of the relationship between export growth and exchange rate change and its volatility. If Pb i¼0ai . 0, then exchange rate depreciation improves exports. If Pc i¼0bi , 0, exchange rate volatility reduces exports through exporters’ responses to perceived risk, while Pc i¼0bi . 0, it stimulates exports. The equation also shows that the net effect relies on the comparison of exchange rate level change effect and exchange rate risk effect. 4.2 Data sources and processing This paper employs bilateral agricultural exports from China to Japan on a monthly basis from January 2002 to April 2007. Seasonally adjusted real agricultural export revenue with base month January 2002 equals nominal export revenue in US dollar deflated by the consumer price index (CPI) of the US and the export price index. CPI comes from the International Financial Statistics (IFS) and Direction of Trade