Faculty Economics and Business
The effect of a Chinese renminbi appreciation
on the US-China trade imbalance
By: Michalli Harmsen
Student number: 10004560
Bachelor thesis
2013-2014
1
Table of contents
1 Introduction
2
2 Theory
3
2.1 Elasticities approach and absorption approach
4
3 Literature Review
6
3.1 Positive relation between imbalance and renminbi
6
2.2 Doubtful effectivity of a RMB appreciation
7
4 Methodology and Data
10
4.1 Methodology
10
4.2 Data
14
5 Results
16
5.1 Base model
16
5.2 Specific sector models
18
6 Discussion and Conclusion
22
2
1 Introduction
During the past twenty years the US has grown a larger amount of external debt than any other country. By the end of 2012, this amount exceeded 16 trillion US dollars (Treasury Direct, 2013). Furthermore, external debt had grown by over a trillion USD per year the five years before that. As one of the largest trade partners of the US, China plays an important role in this situation
(United States Census Bureau, 2014a). The US has been running a trade deficit for years. Figure 1 shows how this deficit has been distributed over US trading partners since 1994. Since
2000 China has been the largest contributor to the deficit. In 2013 China accounted for 46% of the deficit. This corresponds to a deficit of almost 314 billion dollars between the US and China.
Zhang explains that “many, if not most” blame the artificially low RMB for the current situation (2012, pp. 171-172). In 1997 China pegged its renminbi against the US dollar at a rate of 8.27 (Thomson Reuters, 2012). The Chinese government kept the USD-RMB exchange constant at this rate until 2005, when it announced it would slowly let the RMB appreciate (Baak, 2008, p. 117). This appreciation was welcomed by the US, since many believe this was needed to improve the US trade deficit (Goldstein, 2007, p. 11). Goldstein claims that a 25% appreciation of the RMB against the USD improves the US current account by 150 billion US dollars and Fair (2010, p. 243) explains that a renminbi appreciation leads to a substitution away from Chinese produced goods towards American goods.
However, others claim that a renminbi appreciation is not an effective method to
improve the trade imbalance. Goldstein and Lardy (2009, p. 56) explain that the price of Chinese imports will not increase due to a large import content of the goods and according to Zhang and Fung (2006, p. 1007) an appreciation might even increase the trade imbalance due to price-inelasticity. This paper aims to find an answer to the question: “to what extent does a Chinese renminbi appreciation reduce the trade imbalance between China and the US?”
In order to answer this question empirical research using multiple regression models is conducted. First the effect of an appreciation on total trade between both countries is
3
data from the period 1994-2013, which allows for both the full period of the RMB-USD peg as well as two periods of appreciation to be taken into account.
This paper is structured as follows. In Section 2 economic theory concerning exchange rates and the trade balance is discussed. Section 3 provides an overview of the results of research regarding US-Sino trade and the USD-RMB exchange rate. Section 4 presents the methodology and describes and summarizes used data. The results of this research are illustrated and interpreted in Section 4 and finally in Section 5 an answer to the research question is given.
2 Theory
In this section dominating theories concerning international trade and exchange rates are discussed. The two most important theories concerning the effect of exchange rates on
Figure 1. United States Census Bureau. (2014a). Top trading partners. Retrieved from http://www.census.gov/foreign -trade/statistics/highlights/top/top1312yr.html
Note: figures computed by author using data from the above source
-10% 0% 10% 20% 30% 40% 50% pe rce nt age of t ot al t rade de fi ci t
Figure 1. US trade deficit by country
Japan China Malaysia Saudi Arabia Italy Germany Ireland Venezuela Mexico Canada Others
4
a country’s trade balance are the elasticities approach theory and the absorption approach theory. The two theories are described consecutively.
2.1 Elasticities approach and absorption approach
Lerner (1944, p. 377) describes how a domestic currency depreciation has two effects on the trade balance, which each work in opposite directions. When the currency depreciates, foreign imports become more expensive and exports become cheaper. So the depreciation affects the prices of imports and exports. This price effect has a negative influence on the trade balance, he explains. However, there is a second effect that positively alters the trade balance, Lerner points out. When imports become more expensive, its quantity decreases. Likewise, foreign countries will demand more of the home country’s exports. This quantity effect has a positive influence on the trade balance. Which of these effects dominates depends on elasticities regarding exports and imports.
Robinson mentions the foreign elasticity of demand for exports and the home elasticity of demand for imports (1947, p. 138). The former is used when analyzing the export side. As a result of a depreciation, the volume of exports will increase. The more elastic the foreign elasticity of demand for exports, the larger this increase. Since the depreciation does not directly influence the price of export goods expressed in home currency, the price effect is zero (Robinson, 1947, pp. 139-140). On the import side, a depreciation makes imports more
expensive, which increases the value of the imported goods. For simplicity, this price change is assumed to move in proportion to the exchange rate. This means that the price effect is 1. The quantity effect is determined by the home elasticity of demand for imports. A very elastic demand results in a large quantity effect (Robinson, 1947, pp. 141). Combining these effects results in the Marshall-Lerner condition:
5
This condition states that the sum of the foreign elasticity of demand for exports and the home elasticity of demand for imports must be larger than unity, in order for a depreciation to have a positive effect on the trade balance.
Alexander (1952, p. 264) claims that the Marshall-Lerner condition alone is not enough for determining whether or not a depreciation leads to an improvement of the trade balance. He explains that total elasticities instead of partial elasticities should be looked at. The
difference is the following. Partial elasticities evaluate the effect of a depreciation on the trade balance, other things being equal. However, due to a complicated set of relations other things are not equal, he points out. Total elasticities take this into account and therefore depend on the whole economic system. By looking only at partial elasticities, some effects are being left out, which generates an imperfect analysis.
According to Alexander, one main factor that is missing in the elasticity approach theory is domestic absorption. Domestic absorption is defined as the sum of consumption expenditure, government expenditure and investment (1952, p. 265). This results in the following identity.
Where A represents absorption and TB represents trade balance. The identity shows that a country’s trade balance can be defined as the difference between total production within a country and domestic absorption. So when absorption is larger than domestic production the trade balance is negative and when absorption is smaller the trade balance is positive. This shows that domestic absorption in itself influences the trade balance. Moreover, domestic absorption is directly affected by the exchange rate. Alexander continues to set out several direct effects the exchange rate has on absorption.
One example is the cash balance effect (Alexander, 1952, p. 271). It is assumed that a devaluation increases the domestic price level. If individuals wish to hold a constant balance of real money, a price increase leads to a higher money demand, Alexander explains. If the
government does not respond by increasing the money supply, then the only way to satisfy this demand is by cutting down on absorption, he points out. One option would be to sell bonds, for
6
example. This example illustrates that the direct effects of the exchange rate on domestic absorption should not be neglected (1952, pp. 270-275).
3 Literature review
This section describes findings of previous research regarding US-Sino trade and the RMB exchange rate. First articles that claim a positive relation between the trade imbalance and a renminbi appreciation are described. This is followed by papers explaining that a renminbi appreciation might not be an effective way of improving the trade imbalance.
3.1 Positive relation between the imbalance and renminbi
Fair explains that an RMB appreciation could improve the US current account via the substitution effect (2010, p. 243). This implies that when a good becomes more expensive, consumers move to buying the goods substitute. In this case the US substitutes away from Chinese produced goods towards their home produced goods, as prices of Chinese exports increase due to appreciation, Fair explains. This effect on itself would clearly improve the US-China trade imbalance.
However, as Fair also points out, one should not focus solely on the positive effects of a RMB appreciation (2010, p. 235). The second effect Fair names, concerns the fact that a
reduction in Chinese exports also decreases Chinese output. This in turn leads to reduced Chinese imports, partly from the US (2010, p. 243). Third, Fair claims that higher import prices for the US lead to a higher price level overall, which has a negative effect on US output and demand. These two effects show that a RMB appreciation has negative effects on the US-China trade imbalance as well.
Both Fair (2010, pp. 240-244) and Baak (2008, pp. 122-126) aim to quantify these effects using data from the period in which the renminbi was first pegged and then appreciating. Fair finds that, as theory suggests, a RMB appreciation indeed increases the dollar price of Chinese exports (2010, p. 241). On the other hand, since import prices for China decrease, the domestic
7
price level in China decreases. This leads to a decrease in the price of Chinese exports. Overall, the dollar price of Chinese exports increases but this is partly offset by the adjustment of domestic prices, he explains (2010, p. 241). As predicted, this increase in the dollar price of Chinese exports reduces Chinese exports, output and thus imports.
For the US, the higher import price increases the domestic price level, which raises the price of US export goods. The higher price level reduces real wealth and real disposable income. Baak claims that on the other hand, since an appreciation decreases the price of US exports in renminbi, the appreciation also has a positive effect on US exports to China (2008, p. 122).
The above analysis shows that positive and negative effects go both ways. Fairs estimates show that in total, both US exports to China and US imports from China decrease in value, although the decrease in exports is smaller (2010, p. 243). This implies that the effect of a RMB appreciation on the US-China trade imbalance is positive. Baak finds similar results (2008, pp. 118-119). It is important to note here, that the predictions about the size of each effect depend on the assumptions made about the price-elasticities of the demand for exports and imports of both countries, as discussed earlier (Goldstein & Lardy, 2009, pp. 54-55).
3.2 Doubtful effectivity of a RMB appreciation
Between July 2005 and July 2008, the renminbi appreciated 21% against the dollar (Thomson Reuters, 2012). During this period the imbalance kept on growing. In 2005 the imbalance amounted 202 billion USD and by the end of 2008 it increased to 268 billion USD. The
development of the imbalance is shown in Figure 2. Several reasons are provided for the fact that the imbalance grew regardless a possible positive relation between the imbalance and a RMB appreciation.
First, Zhang points out that even though the renminbi appreciated, this was not passed through to the price of Chinese exports to the US (2012, p. 175). Goldstein and Lardy show that when the RMB appreciated 18% during the period June 2005 and March 2008, Chinese export prices to the US increased by only 2.5% (2009, p. 61). They speculate that since the profit margins in Chinese firms did not decrease, this lack of the price pass through is most likely
8
explained by a productivity growth in Chinese industries. Goldstein and Lardy suggest that during 2005-2008 productivity in Chinese industries that import to the US grew by 15% (2009, p. 61). They claim that even though the nominal appreciation of the RMB against the USD was 18%, when one takes this productivity growth into account the RMB actually faced a real depreciation of 15% (2009, p. 61). Therefore, Chinese goods became more competitive during this period.
Goldstein and Lardy provide another reason for failure of the price pass through (2009, p. 56). They explain that the high import content of Chinese export reduces the effectiveness of a RMB appreciation. The average import content of Chinese goods is around 35% (Goldstein & Lardy, 2009, p. 56). A RMB appreciation has the effect that foreign imports become cheaper for China. This means that input costs for Chinese export products, like components and
assemblies decrease. This partially offsets the positive effect of the appreciation on export price (2009, pp. 56-57).
To analyze the influence of the high import content of Chinese exports further, Thorbecke and Smith (2010, pp. 96-98) distinguish between ordinary exports and processed exports. They explain that ordinary exports are labour-intensive goods, produced primarily using domestic inputs. Examples of these goods are toys and textile (China’s Custom Statistics, 2014). Processed goods on the other hand are more capital-intensive. Separate parts are produced in other East-Asian countries and sent to China to be assembled. Examples of these goods are machinery and electrical equipment (China’s Custom Statistics, 2014).
Figure 2. US trade with China
Figure 2. United States Census Bureau. (2014b). Trade in goods with China [Data file]. Retrieved from http://www.census.gov/foreign-trade/balance/c5700.html
9
Thorbecke and Smith investigate the effect of a RMB appreciation against importing countries, keeping the exchange rate between the RMB and the rest of East Asia constant. They find that in this case, a 10% appreciation of the RMB would reduce China’s ordinary exports by 12%, but reduce processed exports only by 4% (2010, p. 96). Since 65% of total Chinese exports to the US in 2006 were processed exports, this limits the effect of a RMB appreciation
(Thorbecke & Smith, 2010, p. 98).
Thorbecke and Smith also explain that a decrease of China’s ordinary exports will not improve the US trade balance (2010, p. 96). Ordinary exports are mainly basic goods, which Europe and the US will simply import from other cheap countries if China’s exports become more expensive, they say. The only way to shift production to Europe and the US would be a generalized appreciation in the entire region (Thorbecke & Smith, 2010, p. 96). Indeed, they find that a general appreciation of East Asian currencies reduces ordinary exports by 12% and processed exports by 10% (2010, pp. 105-106). These results correspond to the opinion of Goldstein and Lardy (2009, p. 66).
Zhang, Fung and Kummer provide another reason for why it is doubtful that a RMB appreciation will reduce the US trade deficit (2006, pp. 45-47). They claim that other factors influence the US trade deficit to a larger extent than the renminbi exchange rate, like a low savings rate and large fiscal deficits. Recently, US trade imbalances with other countries like Saudi Arabia and Finland have also increased greatly (Zhang et al., 2006, p. 46). This shows than one cannot blame the problems the US is facing only on China. Zhang and Fung add to this that since the US trade deficit is caused by other factors, China’s currency revaluation will result in a worsening of the US trade deficit with the rest of the world (2006, p. 1003).
As discussed earlier, if the sum of the price-elasticities of demand for exports and imports is greater than one, a depreciation of the domestic currency will improve a country’s trade balance. Goldstein claims it is likely the Marshall-Lerner condition holds for China, since the price elasticity of demand for China’s exports is high (2007, p. 18). This would mean that a RMB appreciation would indeed lead to a smaller current account surplus for China. According to Zhang and Fung, the opposite is true (2006, p. 1007). They explain that since China exports mainly necessity goods, the price-elasticity of demand for China’s exports is inelastic. In this
10
case, a currency revaluation would lead to an even bigger account surplus for China and trade imbalance between China and the US.
All the above arguments show that it is not clear whether an appreciation of the renminbi decreases or increases the trade imbalance between China and the United States. From 2008 until 2010 the renminbi was kept constant against the USD again, followed by a second period of appreciation during the last decade (Thomson Reuters, 2012). None of the discussed articles includes this second period of appreciation into its dataset. This paper seeks to find an answer to the question of whether or not a revaluation of the renminbi reduces the trade imbalance between China and the US using data from both periods of appreciation.
4 Methodology and data
To assess the impact of an appreciation of the renminbi against the dollar on the trade imbalance between the US and China, three multiple regression models are built. The coefficients are estimated using Ordinary Least Squares. This section first introduces these models after which a summary of the data is presented.
4.1 Methodology
4.1.1 Base model
(1)
Model 1 tests how a renminbi appreciation influences US net exports to China. The dependent variable is the trade imbalance between both countries, defined as US exports to China minus US imports from China. The trade imbalance is expressed in billions of US dollars. To test the impact of a RMB appreciation, two explanatory variables were incorporated into the model. The first is a variable measuring the RMB-USD exchange rate (Srmbusd). The second is a dummy
11
pegged against the US dollar (app). Several control variables were also incorporated into the model.
The first control variable concerns US domestic absorption. To correct for fluctuations in demand US real domestic absorption is incorporated into the model. Real absorption instead of nominal absorption is used to control for the trend of inflation. Values for nominal domestic absorption are found by adding personal consumption expenditure, public consumption expenditure and gross domestic investment. Dividing this by the Consumer Price Index (CPI) yields real domestic absorption. A represents real domestic absorption measured in billions of US dollars.
The second type of control variables concern China’s competitors with respect to trading with the US. As Fair (2010, p. 243) explains, one way exchange rates influence
international trade is through the substitution effect. A depreciation of the domestic currency makes a good cheaper for foreign importers. Therefore, fluctuations in the currencies of China’s competitors should be controlled for. In order to determine who China’s competitors are, the origin and composition of US imports must be examined further.
Figure 3 shows US imports from China per sector for the period 1996-2003. The two biggest sectors are ‘machinery and transport equipment’ and ‘miscellaneous manufactured articles’. Combined these two sectors account for 83% of total imports from China. In both sectors China is the most important trade partner of the US. Figures 4 and 5 display China’s competitors: it shows which other countries the US imports machinery and transport
equipment and miscellaneous manufactured articles from, during 1996-2013. In both sectors Mexico, Canada and Japan account for a large part of imports to the US. When combining all years and both sectors, China accounts for 30% of total imports, followed by Mexico with 15% and Japan with 10%. To control for fluctuations in the competitiveness of these countries, an exchange rate variable for China’s two most important competitors, Mexico (Spesousd) and Japan
(Syenusd), is incorporated into the model.
Important to note here is that the exchange rates are defined as units of foreign currency per US dollar. Therefore, an appreciation of the foreign currency means a lower exchange rate. β1 and β2 are the parameters of interest. If a RMB appreciation has a positive
12
effect on the US-Sino trade imbalance, β1 will be negative and β2 will be positive. If a renminbi
appreciation negatively affects the trade imbalance, β2 will be negative. The opposite will be
the case if a renminbi appreciation negatively affects the trade imbalance.
4.1.2 Specific sector models
(1)
(2)
Model 2 and Model 3 are more specific extensions of the base model. Each of these
models investigates the effect of a Chinese currency appreciation on a separate sector of trade. The independent variables in Model 2 and Model 3 are the same as in Model 1. The dependent variable in Model 2 is the trade imbalance between the US and China of machinery and
Figure 3. United States International Trade Commission. (2014). Interactive tariff and trade dataweb. Retrieved July 16, 2014 from http://dataweb.usitc.gov 0% 10% 20% 30% 40% 50% 60% 70% P e rce nt age of t ot al im ports
Figure 3. US imports from China by sector
Miscellaneous manufactured articles Machinery and transport equipment Manufactured goods classified chiefly by material
Chemicals and related products, n.e.s. Food and live animals
Mineral fuels, lubricants and related materials
Commodities and transactions not classified elsewhere in the SITC
Crude materials, inedible, except fuels Beverages and tobacco
13
United States International Trade Commission. (2014). Interactive tariff and trade dataweb. Retrieved July 10, 2014 from http://dataweb.usitc.gov/
United States International Trade Commission. (2014). Interactive tariff and trade dataweb. Retrieved July 10, 2014 from http://dataweb.usitc.gov/ 0 20 40 60 80 100 120 140 To tal i m p o rts in b ill io n s o f US D
Figure 5. US imports of miscellaneous manufactured articles
China Japan Mexico Canada Taiwan Italy Hong Kong United Kingdom Germany Korea Indonesia Thailand France Dominican Rep Malaysia Others 0 50 100 150 200 250 To tal i m p o rts in b ill io n s o f US D
Figure 4. US imports of machinery and transport equipment
Japan Canada Mexico Germany Taiwan Singapore Korea China Malaysia United Kingdom France Italy Thailand Sweden Philippines Others
14
transport equipment (IMB_Mach). This is defined as US imports of machinery and transport equipment from China minus US exports of machinery and transport equipment to China. In Model 3 this same method is extended to miscellaneous manufactured goods (IMB_Misc).
Figure 3 shows that ‘machinery and transport equipment’ and ‘miscellaneous manufactured articles’ are the two biggest sectors in trade between China and the US. It is interesting to look at these sectors individually, since each sector concerns a different type of goods. First of all, both these sectors consist mainly of processed exports, as classified by Thorbecke and Smith. This is different from Model 1, where all US-Sino imports and exports are taken into account. Second, the two sectors are each different from the other in the type of goods they represent. ‘Machinery and transport equipment’ consists of goods like office and household machinery, televisions and mobile phones, which are normal goods. ‘Miscellaneous manufactured articles’ on the other hand, consists mainly of necessity goods. A full list of the sector classifications is provided by the United States Trade Commission (2014). Zhang and Fung (2006, p. 1007) explain that necessity goods have a lower price-elasticity of demand than normal goods, which means they are less sensitive to price fluctuations. Therefore, each sector might respond differently to a renminbi appreciation. Thorbecke and Smith found that
processed exports respond less strongly to a RMB appreciation than ordinary exports. So it is expected that the parameters of Srmbusd and app are similar in size but smaller in magnitude in
Model 2 and 3 than in Model 1. Also, since Model 3 represents necessity goods, this model is expected to have smaller Srmbusd and app parameters than Model 2.
4.2 Data
The renminbi was floating from 2005 until 2008 and from 2010 onwards. To take both periods of appreciation into account, quarterly data up to 2013 Q4 is used. For Model 1 data is used from the period 1994 Q1 to 2013 Q4. This generates a sample with 80 data points. For Model 2
15
and Model 3 Data is available that goes back to 1996 Q1, which gives a sample with 72 data points. Monthly data about total imports and exports is provided by the United States Census Bureau (2014b). Monthly data about the specific sectors is downloaded from the United States Trade Commission website (2014). From this monthly data quarterly data for the trade balances is computed. Table 2 shows that US net exports to China averaged -40.33 billion US dollars. With a few exceptions, this number grew more negative every period, with a minimum at -90 billion USD. Net exports of machinery and transport equipment and miscellaneous
manufactured articles follow a similar trend but are clearly smaller in magnitude, since they represent only a fraction of total net exports. These variables are almost equal in their average value, but the imbalance of machinery and transport equipment shows a much larger variation.
Data to calculate nominal absorption is provided by the United States Bureau of Economic Analysis (2014), and data about the CPI is provided by the United States Bureau of Labor Statistics (2014). Average domestic absorption equals 16.10 billion USD. Since real instead of nominal absorption was used this variable does not follow a clear trend. The dummy variable averages 0.54, indicating that the RMB is pegged in 43 out of 80 data points.
Data about the exchange rates can be downloaded from Thomson Reuters Datastream (2012). For the exchange rates the value at the end of each period was used. Table 2 to shows that on average 1 US dollar trades for 7.72 yuan, 106.91 yen and 10.10 pesos.
Va ria ble Obs. Mea n Std. Deva tion
IMB 80 -40.3293 26.3264 Srmbusd 80 7.7168 0.8213 app 80 0.5375 0.5017 A 80 16.1002 2.0309 Syenusd 80 106.9058 15.1633 Spesousd 80 10.1014 2.5242 IMB_Mach 72 -20.5610 15.2650 IMB_Misc 72 -20.1656 9.2582
16
5 Results
To examine the effect of the revaluation of the Chinese renminbi on the trade imbalance between China and the United States, several regressions are performed. First the results for Model 1 are presented, followed by the results for the more specific models: Model 2 and Model 3.
5.1 Base model
5.1.1 Results base model
Table 3 shows the results of two variations of Model 1. In the full model (1B), over 94% of the variance of the imbalance is predicted by the variables. Both the dummy and linear variable measuring renminbi appreciation prove to be statistically significant at a 1% significance level. The results show that one unit of renminbi appreciation decreases the trade imbalance with over 7 billion dollars, once controlled for absorption and fluctuations in other important
exchange rates. The dummy variable indicating a period of appreciation confirms this result: on average the imbalance is almost 8 billion dollars lower during a period of appreciation. Spesousd is
not statistically significant, indicating that a Mexican peso appreciation does not significantly affect the trade imbalance. Furthermore, the joint hypothesis that Syenusd and Spesousd equal 0 is
not rejected at a 5% significance level (prob > F = 0.0967), as is seen in the stata output in Figure 6. Excluding Syenusd and Spesousd results in parameter estimates of similar magnitude (1A).
Several tests are performed to test the validity of model 1C. First the normality of the residuals is tested, which seem approximately normally distributed. Second, the model is tested for multicollinearity. Since none of the estimators show a variance inflation factor (vif) higher than 10 it can be assumed that the model does not suffer from multicollinearity. Third, the model is tested for heteroskedasticity. The Breusch-Pagan / Cook-Weisberg-test reports a p-value of 0.0027, indicating heteroskedasticity of the standard errors. Repeating the regression using robust standard errors generates similar results, as is seen in Table 1.
17
5.1.2 Interpretation of results Model 1
The results of Model 1 show that the exchange rate is positively correlated with the trade imbalance: when the renminbi appreciates, and thus the exchange rate falls, the trade imbalance increases in magnitude. This means that when the renminbi becomes stronger against the dollar, either the value of American exports decreases or the value of Chinese imports increases, or both. These results contradict with the findings of Fair (2010, p. 243) and Baak (2008, pp. 118-119), who both find that a renminbi appreciation should have a small but positive effect on the trade imbalance. This difference is likely to be caused by the difference in datasets. Fair and Baak use data ranging from 1999 to 2008 and 1991 to 2006 respectively, while this paper uses a larger and more recent dataset from the period 1994-2013. This allows for the appreciation period after 2010 to be taken into account as well. If the difference in datasets is the cause for the contradicting results, this means that during the second period of appreciation the imbalance responded differently to renminbi fluctuations than during the first period. The results in this paper do not provide an explanation for this. However, it would be interesting to research this in the context of the productivity growth mentioned by Goldstein and Lardy (2009, p. 61). If the price pass through of a RMB appreciation was hindered by a productivity surge during 2005-2008, analyzing these effects for the period after 2010 should provide more insight in the situation.
The finding that Srmbusd and IMB have a positive relation confirms Zhang and Fung’s
(2006, p. 1007) claim that a renminbi appreciation has a counterproductive effect on the trade imbalance. Their explanation for this claim is that the demand for goods that the US imports from China is relatively price-inelastic, allowing the price effect of the appreciation to dominate the quantity effect. In this context, it is not surprising that the joint contribution of Syenusd and
Spesousd is not statistically significant. These variables represent price fluctuations of competing
producers. If US imports from China have a price-inelastic demand, price changes of substitutes are not relevant either. However, this paper cannot make any conclusions on the price and quantity effect of the appreciation, since only the impact of the appreciation on net exports is
18
examined. Future research examining the effect of a RMB appreciation on US imports and exports separately, could provide more information on the magnitude of the price and the quantity effect.
The fact that absorption has a negative effect on the trade imbalance corresponds to Alexanders theory (1952, p. 265). He shows that Y – A = X – M. This means that when A increases, net exports decrease. The estimates of Model 1 show that one unit increase in absorption reduce the imbalance with 10.03 units.
Table 3. Results for regression of general trade imbalance
Dependent variable: General trade imbalance between US and China
1A 1B Coefficient (Std. error) t-statistic Robust t-statistic (Std. error) Coefficient (Std. error) t-statistic Robust t-statistic (Std. error) Srmbusd 8.7963*** (1.1873) 7,41 4.19 (2.0971) 7.3844*** (1.4056) 5,25 2.76 (2.6716) app -9.0083*** (1.5489) -5.82 -5.56 (1.6189) -7.8857*** (1.7921) -4,40 -3.46 (2.2780) A -9.6571*** (0.4538) -21.28 -13.08 (0.7383) -10.0268*** (8.2101) -12,21 -11.35 (0.8832) Syenusd 0.1371** (0.0649) 2,11 2.04 (0.0671) Spesousd 0.2896 (0.7099) 0,41 0.27 (1.0677) Intercept 52.1151*** 3,42 1.87 (27.7982) 50.7807 (15.0014) 3,39 2.10 (24.1595) No. Of observations 80 80 R^2 0,9440 0,9472 Prob > F; F-statistic 0,0000 427,23 415,17 0,0000 265,58 264,8
Note:***, ** or * indicates a significance level of 1%, 5% or 10% respectively.
5.2 Specific sector models
19
The same regression is run for the sectors ‘miscellaneous manufactured articles’ and
‘machinery and transport equipment’. The results for these regressions are shown in Table 4 and Table 5. For both models robust standard errors are used. The results for the models concerning individual sectors differ from the results for the base model on a few points. Model 2 is discussed first.
When comparing the estimates of Model 2 to Model 1, it is seen that the coefficients have the same sign but are different magnitude. When looking Model 1A and 2A, the
parameters for the variables app and A are roughly twice the size of their Model 2
counterparts, while the parameter for Srmbusd is nearly the same. On average net exports are
4.75 billion USD smaller during a period of a floating RMB in Model 2A, compared to 9.01 billion USD in Model 1A. Again, the joint hypothesis for excluding Syenusd and Spesousd from the model is
rejected at a 5% level, as Figure 6 shows. This model too provides a good fit for explaining the variation in the data, shown by an R^2 of 0.9472. The model is tested for multicollinearity and normality of the residuals. None of the variables shows a vif-value higher than 10 and the residuals are approximately normally distributed.
The results for Model 3 are presented in Table 5. In the full model, neither one of the foreign exchange rate variables is statistically relevant and the hypothesis that these variables jointly contribute to the model is rejected. When looking at the results for the regression excluding these exchange rates (3A), all remaining explanatory variables are statistically significant. This model shows an R^2 of 0.8945, indicating a good fit of the model. When comparing the coefficients with Model 1 and 2, it is clear that the estimates are much smaller here. While 1 unit of appreciation decreases the trade balance with 8.80 billion dollars in Model 1 and 9.27 billion dollars in Model 2, in Model 3 this number is 3.83 billion. This indicates that the trade imbalance of miscellaneous manufactured articles is less sensitive to price
fluctuations. This model too shows no sign of multicollinearity and has approximately normally distributed residuals.
20
As is Model 1, an appreciation of the RMB worsens net exports of machinery and transport equipment and miscellaneous manufactured articles. However, this effect is smaller in Model 2 and even smaller in Model 3. In the light of Thorbecke and Smiths (2010, p. 96) results, this is not surprising. They explained that due to a high import content in processed exports, the price pass through effect of a RMB appreciation is smaller compared to ordinary exports. Since Model 2 and 3 concern solely processed exports, the estimates of Srmbusd and app are expected
to be smaller than in Model 1.
It is also interesting to look at the differences between Model 2 and Model 3, since Model 2 concerns normal goods and Model 3 concern necessity goods. Zhang and Fung explain that due to a lower price-elasticity of demand, price fluctuations of necessity goods have a smaller effect on demand than with normal goods. This means that the coefficients of Srmbusd
and app should be smaller here than in the other two models, which is the case indeed. Notable is that the effect of absorption on the dependent variable in the two specific models is a lot smaller than in the base model. Higher absorption worsens the trade balance by 4.3 units in Model 2 and 3.7 units in Model 3, while in Model 1 this effect is 10.03 units. It seems plausible that the reason for this is that when absorption increases, the effect of this increase is spread out over all sectors. So when looking at only one sector, this effect seems smaller.
Table 4. Results for regression of ‘machinery and transport equipment’ trade imbalance Dependent variable: trade imbalance of machinery and transport
equipment between US and China
2A 2B Coefficient (Std. error) t-statistic Coefficient (Std. error) t-statistic Srmbusd 9.2745*** (0.6672) 13,90 7.8813*** (0.8949) 8,81 app -4.7532*** ( 0.6881) -6.91 -4.9835*** (0.7104) -7,01 A -4.5195*** (0.2172) -20.81 -4.3441*** (0.4486) -9,68 Syenusd 0.0630* (0.0340) 1,85
21 Spesousd -0.3433 (0.6406) -0,54 Intercept -14.8139* (8.1701) -1.81 -10.0360 (9.048) -1,11 No. Of observations 72 72 R^2 0,9681 0,9703 Prob > F; F-statistic 0,0000 754,86 0,0000 459,52
Note:***, ** or * indicates a significance level of 1%, 5% or 10% respectively.
Table 5. Results for regression of ‘miscellaneous manufactured articles’ trade imbalance Dependent variable: trade imbalance of miscellaneous manufactured
articles between US and China
3A 3B Coefficient (Std. error) t-statistic Coefficient (Std. error) t-statistic Srmbusd 3.8291*** (0.7847) 4,88 4.5402*** (1.2063) 3,76 app -2.0135** (0.8086) -2.49 -1.6747 (1.1308) -1,48 A -3.4520*** (0.2281) -15.13 -3.6921*** (0.5104) -7,23 Syenusd -0.0135 (0.0363) -0,37 Spesousd 0.4254 (0.7921) 0,54 Intercept 8,3996 0,90 3.6452 (13.0915) 0,28 No. Of observations 72 72 R^2 0,8945 0,8960 Prob > F; F-statistic 0,0000 242,29 0,0000 152,21
Note:***, ** or * indicates a significance level of 1%, 5% or 10% respectively.
Figure 6. Stata ouput for testing joint hypothesis of foreign exchange variables
Model 1 Model 2 Model 3
Prob > F = 0.0967 F( 2, 74) = 2.41 ( 2) Spesousd = 0
( 1) Syenusd = 0 . test Syenusd Spesousd
Prob > F = 0.1077 F( 2, 66) = 2.30 ( 2) Spesousd = 0
( 1) Syenusd = 0 . test Syenusd Spesousd
Prob > F = 0.7359 F( 2, 66) = 0.31 ( 2) Spesousd = 0
( 1) Syenusd = 0 . test Syenusd Spesousd
22
6 Conclusion
This paper aimed to determine the effect of a Chinese renminbi appreciation on the trade imbalance between the US and China. In order to do this, several linear regression models were established. The first model concerned total net exports between both countries and two extensions of Model 1 looked at the trade imbalance of machinery and transport equipment (Model 2) and miscellaneous manufactured articles (Model 3). To isolate the influence of a renminbi appreciation, US domestic absorption, the peso exchange rate and the dollar-yen exchange rate were controlled for.
The results of Model 1 show that contrary to what was previously found, a renminbi appreciation worsens the trade imbalance between the US and China. This is the case for the total trade imbalance and the trade imbalances of the two biggest sectors in US-Sino trade. Furthermore, it was found that the effect of the appreciation is smaller in the sectors representing processed exports. This can be attributed to the high import content of export goods in these sectors. Finally, the influence of a renminbi appreciation is smaller in the sector consisting of necessity goods than in the sector consisting of normal goods. This is likely to be caused by an inelastic price-elasticity of demand for necessity goods.
To conclude, this paper provides evidence for rejecting the hypothesis that a Chinese renminbi appreciation is an effective way to improve the trade imbalance between China and the US. If both countries improve this imbalance, a different channel via which to this should be looked for.
23
7 Reference list
Alexander, S. S. (1952). Effects of a devaluation on a trade balance. International Monetary Fund Staff Papers, 2(2), 263-278
China’s Custom Statistics. (2014). Trade statistics. Retrieved August 14, 2014 from http://www .customs-info.com/Trade/Commodity.aspx
Baak, S. (2008). The bilateral real exchange rates and trade between China and the U.S. China Economic Review, 19(2), 117-127
Fair, R. C. (2010). Estimated macroeconomic effects of a Chinese Yuan appreciation (Working paper No. 1755). Retrieved June 11, 2014 from http://cowles.econ.yale.edu/P/cd /d17b/d1755.pdf
Goldstein, M. (2007). A (lack of) progress report on China’s exchange rate policies (Working paper May 2007). Retrieved June 20, 2014 from http://www.iie.com/publications /papers/goldstein0507.pdf
Goldstein, M., & Lardy, N. (2009). The future of China’s exchange rate policy. Washington, D. C.: Peterson Institute for International Economics
Lerner, A. P. (1944). The economics of control: principles of welfare economics. New York: The Macmillan Company
Robinson, J. (1947). Essays in the theory of employment. Oxford: Basil Blackwell
Thomson Reuters Datastream. (2012). Chinese Yuan to US dollar – exchange rate. New York: Thomson Reuters
Thorbecke, W., & Smith, G. (2010). How Would an Appreciation of the RMB and Other East Asian Currencies Affect China’s Exports? Review of International Economics, 18(1), 95-108
24
Treasury Direct. (2013). Historical debt outstanding. Retrieved June 24, 2014 from http ://www.treasurydirect.gov/govt/reports/pd/histdebt/histdebt_histo5.htm
United States Bureau of Economic Analysis. (2014). National income and product account Tables. Retrieved July 10, 2014 from http://www.bea.gov/iTable/iTable.cfm?reqid =9&step=1&acrdn=2#reqid=9&step=1&isuri=1&903=82
United States Bureau of Labor Statistics (2014). Consumer price index – all urban consumers. Retrieved July 24, 2014 from http://data.bls.gov/cgi-bin/surveymost
United States Census Bureau. (2014a). Top trading partners. Retrieved July 11, 2014 from http ://www.census.gov/foreign-trade/statistics/highlights/top/top1312yr.html
United States Census Bureau. (2014b). Trade in goods with China [Data file]. Retrieved June 15, 2014 from http://www.census.gov/foreign-trade/balance/c5700.html
United States International Trade Commission. (2014). Interactive tariff and trade dataweb. Retrieved July 10, 2014 from http://dataweb.usitc.gov/
Zhang, J., & Fung, H. (2006). Winners and losers: Assessing the impact of Chinese Yuan appreciation. Journal of Policy Modeling, 28(9), 995-1009
Zhang, J., Fung, H., & Kummer, D. (2006). Can renminbi appreciation reduce the US trade deficit? China & World Economy, 14(1), 44-56
Zhang, X. J. (2012). Will RMB appreciation reduce trade deficit in the US? Journal of Asia Pacific Economy, 17, 171-187
