1
University of Amsterdam
China’s exchange rate policy, determining the
possible undervaluation of the Renminbi.
Faculty: Faculty of Economics and Business Study Program: BSc Economics and Business Specialization: Economics and Finance
Student Name: Spike Gontscharoff Student Number: 10207007 Supervisor: Stephanie Chan
2 Abstract
The RMB has been the subject of controversy for a long period of time and is currently under pressure of revaluation. The aim of this paper is to determine the equilibrium value of the Chinese exchange rate, using a BEER approach by Kim and Korhonen (2005) and to estimate if the RMB is indeed undervalued. Four factors have been regressed on the real exchange rate (RER) and the real effective exchange rate (REER), using an OLS regression. The OLS estimates show that except for the variable gdpt, the size and sign of the coefficients are as explained by the theory. However, the OLS estimates
are not significant, probably due to the lack of data and the presence of unit roots and serial
correlation. A BEER approach using the RER, estimates the RMB has been overvalued between 1978 and 1984 and mostly undervalued for the period 1985-2012. The BEER approach using the REER finds a contradicting result for the period 2003-2010, in which the RMB was overvalued.
3
Table of contents
Section 1: Introduction 4
Section 2: Development exchange rate regimes 5
Section 3: Signs of undervaluation 6
Section 3.1 Large surpluses both on the current and the financial account 6 Section 3.2 The rapid increase of foreign exchange reserves 7
Section 4: Effects of a possible RMB misalignment 8
Section 4.1 Possible effects for the Global economy 8 Section 4.2 Possible effects for the Chinese economy 9
Section 5: Literature review 10
Section 5.1 The Purchasing Power Parity Approach (PPP) 10 Section 5.2 Macroeconomic Balance (MB)/
Fundamental equilibrium exchange rate (FEER) 11 Section 5.3 Behavioral Equilibrium Exchange Rate (BEER) model 12
Section 6: Data 13
Section 7: Method 14
Section 7.1 Model 14
Section 7.2 Pre-regression tests 15
Section 7.3 OLS pre-estimation tests 15
Section 8: Hypothesis 16
Section 8.1 predicted coefficients RER 16
Section 8.2 predicted coefficients REER 17
Section 9: Results 18
Section 9.1 Results RER 18
Section 9.2 Results REER 19
Section 9.3 misalignment of the RMB, RER 19
Section 9.4 misalignment of the RMB, REER 20
Section 10: Conclusion and Discussion 20
Section 10.1 Discussion 20
Section 10.2 Conclusion 21
Reference list 23
4 Section 1: Introduction
In the last few decades the Chinese economy has been the subject of extraordinary economic growth. Its GDP per capita has increased from 193 U.S. dollars per capita in 1980, to a GDP per capita of 6091 U.S. dollars in 2012 1. The Chinese economy is currently among the largest in the world but is also being criticized by other countries. A large part of this criticism is aimed at China’s exchange rate policy and the possible undervaluation of its currency the Renminbi (RMB) 2.
The RMB has been the subject of controversy for a long period of time and is currently under pressure of revaluation. According to Chang (2007), an increasing number of foreign officials, businessmen and economist, in particular those from Japan and the United States are calling for such a revaluation. They fear that the long term undervaluation has given the Chinese economy an unfair comparative advantage in international trade. The opinions on a possible Renminbi appreciation differ however. Chinese economist for example point out the position of China in the 1997 Asian financial crisis. They show that while surrounding countries experienced a depreciation, the RMB maintained its value, effectively having an appreciation of the Chinese currency. Moreover Chang (2007), mentions that a quick and uncontrolled revaluation could lead to destabilization of the Chinese and potentially the global economy.
China is not the only country whose exchange rate regime has been under investigation. Research by Frankel and Wei (2007) shows that ‘’exchange rate regimes in emerging markets have been the primary concern of international economists and policy makers since the 1990’s’’. The reason for the extra attention on China comes from the fact that it stands out. ‘’China is by far the largest emerging economy that has maintained a pegged currency since the Argentine and Turkish crisis of 2001’’ . The aim of this paper is to actively contribute to the discussion on the Chinese exchange rate. This paper tries to estimate the equilibrium exchange rate for the RMB using a regression model based on four carefully selected independent variables, according to a BEER approach. The regression analysis is based on the ideas by Kim and Korhonen (2005). In this paper I use the basic ideas from the paper regarding the chosen variables and apply them in trying to determine the potential undervaluation of the RMB for the period 1978-2012 using the real exchange rate and for the period 1980-2012 using the real effective exchange rate.
In order to properly asses the research question this paper is organized in 10 Sections. In the second Section an overview of developments in the Chinese exchange rate is presented. Section 3 looks at the possible signs for a RMB undervaluation, while Section 4 covers the effects of a possible
under/overvaluation. In Section 5 a literature review is given to show the different methods used and the contradicting results of these methods. Section 6 and 7 describes the data and the method used. Section 8 and 9 contain the hypothesis and the actual results. Section 10 contains a discussion and conclusion.
1 Source: WDI, http://databank.worldbank.org/data/views/reports/tableview.aspx
2 Yuan is the actual currency unit in China, although the definition Renminbi (RMB) is used by Chinese officials and in most of
5 Section 2: Development exchange rate regimes
Before calculating the equilibrium value of the RMB and the possible under/overvaluation , it is important to get a thorough understanding of the Chinese exchange rate. To get this understanding we look at exchange rate regimes, policy adjustments and the value of the RMB over time.
The Chinese exchange rate regime has undergone several phases. Until 1978 China’s economy was closed. As a consequence there was little to no room for the exchange rate as a part of the foreign trade system. With the start of the economic reform in 1978, policymakers indeed recognized the value of the exchange rate as an instrument for regulating the external sector. In August 1979 the Chinese government decided to adopt a dual exchange rate system that eventually went into effect in 1981.This dual exchange rate system consisted of an internal settlement rate for trade transactions and an official rate for nontrade transactions. The internal settlement rate was set at 2,80 Yuan per U.S. dollar for the period 1981-1984. During this same period the official rate
fluctuated between 1,705 and 2,327 Yuan per dollar.
The period between 1985 and 1990 consisted of several large devaluations of the RMB compared to the dollar (Chou and Shih, 1997).This first devaluation took place in 1986, which devalued the currency from RMB 3,20 per U.S. dollar in 1985 to a value of RMB 3,72 in 1986. The rate was further devalued to RMB 4,72 per U.S. dollar in 1989 and eventually brought down to RMB 5,22 in 1990. In 1991 the Chinese government took its first steps in establishing a more open
economy with the introduction of a managed float system. By allowing the RMB to adjust to market conditions in small regulated steps, the currency further devalued to RMB 5,80 in the period from 1991 till 1993 (Chou and Shih 1997).
A big change took place in January 1994 when China started to further reform its foreign exchange rate system. China’s dual exchange rate system using the internal settlement rate and the official rate was unified into a single exchange market (Bosworth 2004). The unification of the exchange rate market had two consequences. One, it created an interbank market for trading in foreign exchange. Two, it had a substantial impact on this new unified currency, further devaluating it to RMB 8,3 per U.S. dollar. In 1995 the People’s Bank of China (PBC) was legally given the status of central bank by the National People’s Congress3. Although the PBC defines the period from 1995-2005 as a managed float, in reality the RMB could be seen as being pegged to the dollar during this period, maintaining its value of around RMB 8,3 per U.S. dollar. These developments can be seen in Figure 1.
In July 2005, China announced the switch to a new exchange rate regime. This new regime started with a minor revaluation and pegged the Renminbi to a basket of currencies instead of only to the dollar. Also the currency was allowed a movement of ± 0,3 % a day. Frankel and Wei (2007 ) mention that although this new policy allowed for a maximum revaluation of 6,4 % per month, actual revaluation to the dollar was only 6% by the end of 2006. For comparison, the movement of the dollar against other major currencies like the euro and the yen was much larger during this same period.
Frankel and Wei (2007) do stress that because the composition of this basket of currencies is not exactly known, one has to be careful in determining the changes of the exchange rate compared to a specific currency. To cope with these problems they perform a regression analysis to determine the exact composition of the currency basket. They find that most weight is still being put on the dollar but ‘’ subsequently there has been a modest but steady increase in flexibility with some weight shifted to a few non-dollar currencies, but not those one might expect’’. Figure 2 shows that since 2005 the RMB/USD exchange rate has gradually been appreciating, reaching a value of around RMB 6,02 per dollar near the end of 2013.
6
A quick glance at why China is the largest developing economy to maintain a pegged currency looks at two important macroeconomic concepts, economic growth and price stability. The Chinese exchange rate regime can be regarded as successful ,particularly in terms of these two concepts. Figure 3 clearly states the facts, GDP growth was above 7,5 % at all time and inflation was almost always below 5% and remaining fairly stable.
Section 3: Signs of undervaluation
As mentioned before by Chang (2007), China is being put under pressure by large trading partners to allow for RMB appreciation. It could be helpful to determine the foundations for these acquisitions. Arguments for an appreciation are mostly based on the Chinese balance of payments surplus and the rapid increase of foreign exchange reserves at the PBC.
3.1 Large surpluses both on the current and the financial account
The balance of payments (BoP) is a method used by countries to monitor all international transactions at a specific period in time. Originally this BoP was divided into three subaccounts, BoP = CA + CP +OR. The current account (CA) which records cross-border transactions in goods and services. The capital account (CP) records private financial transactions. The last part is the official reserves account (OR) which measures a central banks foreign exchange reserves.
However due to the current initiative of institution like the OECD and the IMF, new definitions are used. The capital account now only measures the transfer of ownership of assets, so-called wealth transfers4 . Also the original capital account is combined with the official reserves account to form the financial account (FA).This gives the new identity: BoP = CA + CP +FA5 (Gartner 2005). Because the new capital account only makes up for only a very small portion of the BoP identity, this subsection only takes a look at the current and the financial account.
Every transaction that involves a purchase of domestic currency is a positive item on the BoP and every purchase involving the sale of domestic currency is a negative item. Purchase of domestic currency is only possible if someone else is willing to sell. This means that the foreign exchange market always is in equilibrium and why the identity for the BoP can be used to describe the foreign exchange market.
According to the new BoP identity, the current account should be balanced against the combined capital and financial account. This means that without intervention from a country’s central bank a current account surplus should be accompanied by a deficit on the capital account and/or financial account. China has maintained a surplus on its current account for a long period of time. Having a current account surplus for multiple years as an emerging country is atypical according to Coudert & couharde (2007). Even more unusual is that China also has a substantial surplus on the financial account. The theory suggests that if China is to restore its balance of payments the RMB should appreciate. Figure 4 gives a summary for China’s Balance of Payments, where it becomes clear that this phenomenon is indeed occurring. Although these numbers on the country’s current and financial account suggest a RMB appreciation, in the case of China they do not justify this entirely. The reason for this is that a number of so called China specific characteristics play an important role that may justify these imbalances instead of a undervalued RMB.
The first being the composition of the Chinese current account. The main part of the imbalance on the current account is caused by the inequality of exports and imports. As can be seen
4 http://www.imf.org/external/pubs/ft/bop/2011/pdf/chapter15.pdf
5 The reason this is explicitly mentioned is the usage of different terminology in papers and the fact that the two methods are
7
from Table 1, exports and imports as percentage of GDP are increasing but with exports increasing at a higher rate. US officials see these low export prices and high export volumes as a result of a RMB undervaluation. China’s domestic production of goods however consist of a small number of products, Table 2. These low export prices could also be the result of an increased cost efficiency due to mass production of certain goods and lower wages in China. Bosworth (2004) further mentions an important point regarding these accusations from American officials. The US-China trade balance shows a large bilateral trade surplus with the United States, possibly indicating a RMB
undervaluation. America on the other hand has bilateral trade deficit with a lot of countries and ‘’China is only one of many countries with which the United States has a trade deficit’’. When looking for the appropriate exchange rate it is China’s overall balance with the world economy that matters. From this perspective, a trade surplus with the US is offset by trade deficits with countries like Japan and other major Asian economies.
A second alternative explanation for the high surpluses on the current account is the savings rate in China. The Chinese saving rate is among the highest in the world. Since 2005 gross national saving has exceeded 50 % of GDP ,see Table 3. There are a number of reasons for this high savings rate. Chamon & Prasad (2010) mention the most important determinants. Growth of per capita GDP has enabled more Chinese to save a part of their money, there are limited possibilities in the country itself to obtain a loan and due to rising macroeconomic uncertainty in China more precautionary savings take place. While national saving rates are rising, investment opportunities in China are still limited. In recent years, China’s financial market has undergone some significant changes, still it remains underdeveloped compared to most industrialized countries. This underdevelopment limits the possibilities for investing. Another way to look at the trade identity of an open economy is the
following : Current Account = Exports – Imports = Savings – Investment (Krugman, Obstfelt & Melitz, 2011). These high savings combined with limited investing could be the cause of a current account surplus, instead of being explained by an undervalued RMB.
3.2 The rapid increase of foreign exchange reserves
There are two aspects regarding this phenomenon . The first one being pointed out by Coudert & Couharde (2007). They mention the accumulation of foreign reserves by the PBC has accelerated since the beginning of 2000. In 2012 these reserves reached 3311,5 billion U.S. dollar making them easily the largest foreign exchange reserves in the world ,Table 4.These foreign exchange reserves are for some part the result of repeated intervention by the PBC, therefore keeping the RMB below its equilibrium level. Due to the high demand for Chinese exports, demand for the RMB started to increase. To prevent the RMB from appreciating to rapidly the PBC sold the RMB to buy U.S. dollars. In theory there is an adjustment mechanism on the exchange market as a reaction to the accumulation of foreign reserves. Holding these amounts of foreign reserves would increase the domestic monetary aggregates. Table 4 shows that indeed monetary aggregates (M2) have been very high in the last two decades. An increase in M2 should cause inflation thereby increasing the real exchange rate6 and causing appreciation of the domestic currency. In China effects of this adjustment mechanism have been limited. Coudert & Couharde (2007) show two reasons for this. One, part of
6 Real exchange rate (rer) is defined as P*/ (E x P). Where P* is the foreign price level, E is the nominal exchange rate (in
8
the effect is reduced due to sterilization7 actions by the PBC. Two , Table 6 shows that even with China’s high growth rate inflation as measured by CPI has remained very low.
A second aspect regarding the rapid increase of foreign exchange reserves is the composition of these reserves. China has maintained a trade surplus for a long period of time now. The country’s foreign reserves are even bigger than this trade surplus, meaning there is capital inflow coming from other sources as argued by Chang (2007). A part of this capital inflow can be explained by the large amount of FDI. Even with this FDI, not all foreign exchange reserves are accounted for. Goldstein (2004) argues that the remaining part of the foreign reserves came from other legal /semi-legal channels. But what is more important, is that these extra capital inflows are there because of speculation on a future RMB revaluation, indicating that the market believes the RMB is undervalued.
Section 4: Effects of a possible RMB misalignment
China’s exchange rate policy has been under a lot of attention. Because the Chinese economy is growing at a rapid rate its influence on the global economy is getting bigger as well. To determine why so much attention is given to the determination of the equilibrium exchange rate ,this sector takes a look at the economic consequences from a under/over valuation. This sector can be further separated into two subsectors:
4.1 Possible effects for the Global economy:
Two of the biggest advocates asking for a RMB appreciation are Japan and the U.S. In February 2003 in a meeting of the G-7 8 , Japan openly demanded an appreciation of the RMB. Goldstein and Lardy (2006) point out that ‘’the long term intervention on the foreign exchange market by the PBC combined with the running current account and trade surplus increase the possibility of protectionist measures by the U.S. and other major industrialized countries’’. These protectionist measures take on different forms. The most simple one is a tax levied on imported goods, called a tariff. Nowadays, non-tariff measures like import quotas and export restraints are getting more important. The majority of economists believe protectionist measures like these are welfare decreasing. In the worst case two trading countries keep implementing these protectionist measures resulting in so called trade war (Krugman, Obstfelt & Melitz, 2011). A trade war can best be explained by using an example. In the case that trade is viewed as a single trade relationship between China and another foreign country and countries have two choices: free trade or protection, payoffs are given in Table 7. Without trade agreements both countries would be better of implementing the strategy protection given they know the strategy from the other country. This results in a so called Nash-equilibrium. Payoffs for both countries however could be higher if both were able to implement the strategy of free trade. Although this is a simplified example it reflects the possible consequences of protectionist behavior.
Another problem due to the growing economic importance of China is its increasing importance on macroeconomic misalignment in the global economy. Preceding the recent financial crisis a lot of economies experienced a current account imbalance. If the RMB truly deviates from its
7 A form of monetary action in which a central bank seeks to limit the effect of inflows and outflows of capital on the money supply. Sterilization most frequently involves the purchase or sale of financial assets by a central bank, and is designed to offset the effect of foreign exchange intervention. Source: http://www.investopedia.com/terms/s/sterilization.asp
8The G-7 was formed in 1975 and initially comprised six nations - France, Germany, Italy, Japan, the U.S. and U.K. - with Canada invited to join the group in 1976. G-7 officials meet periodically to discuss international economic and monetary issues, source:http://www.investopedia.com/terms/g/g7.asp
9
equilibrium level and continues to do so in the future, then the possibility of a macroeconomic misalignment in other countries increases as well. Take for example the U.S. current account deficit. U.S. officials claim this is due to the under appreciation of the RMB. Since 2005, the U.S. current account deficit as percentage of GDP has been decreasing, Table 8. If however these imbalances were to increase again in the future, market uncertainty could grow, causing severe problems on the U.S. market. These effects could spill over to other countries and even possibly cause another crisis (Rodrik 2010).
4.2 Possible effects for the Chinese economy
Having a fixed exchange rate regime limits the possibility of implementing monetary policy, this is no different for the Chinese economy. Officially the RMB is pegged to a basket of currencies and is described by the PBC as a managed float. Goldstein and Lardy (2006) argue that during periods of high investment growth as experienced by China in recent years, demand for domestic loans goes up, increasing the nominal lending rate. The PBC of China however has been reluctant to actually allow for such an increase in the interest rate. They fear that a higher interest rate would also attract more FDI, increasing the demand for the RMB and eventually causing an appreciation. As a result, supply for domestic loans did not keep up with the demand. Policies like this reduce the overall welfare by limiting the amount of value adding activities that can be undertaken.
A second problem China is facing as a result of a possible future appreciation is described in the previous Section, China has been accumulating foreign reserves at an increasing rate. In 2012 these foreign exchange reserves amounted to 3.311,589 billion U.S. dollar ,Table 4. In the same year China’s GDP denoted in U.S. dollar was 8.227,103 billion U.S. dollar9
. This means that in 2012 foreign reserves were 40,25 % of GDP. Some articles argue that the RMB could be undervalued by as much as 20%. If a 20% appreciation actually took place , a capital loss of around 8% of GDP would be the result. With numbers like these reluctance of the PBC for a sudden appreciation of the RMB seems more logical.
A third problem for China is the competiveness of its export sector. Chinese officials fear that appreciation of the RMB would diminish the sale of export goods, reducing economic growth and causing unemployment to rise. This problem is particularly relevant for China because of the composition of its exports, the problem is illustrated by Table 2. China’s export is shifting towards a higher percentage of manufactured goods. In particular the machinery sector has become very large in a short amount of time. Amati & Freund (2010) point out that Chinese exports are highly concentrated in a small number of goods. This seems logical from a trade theory point of view, by focusing on just a few goods, maximum cost efficiency can be achieved due to external and internal economies of scale. Still this strategy becomes a potential problem in the case of a currency appreciation. In the current global economy differences between the production possibilities are getting smaller. One of the reasons China has been able to maintain its edge in the export market is due to its low cost strategy. A large currency appreciation could have substantial effects on this obtained production advantage. Supporters of the RMB appreciation, mention that China could tackle this problem if they were to focus more on the domestic market for its products. According to the PBC however,
developing an internal market this big takes time and cannot be realized immediately.
10 Section 5: Literature review
There has already been done a lot of research regarding the determination of the equilibrium exchange rate. The estimation of the Chinese equilibrium exchange rate has been performed by previous
articles, however findings of these studies vary a lot and are quite contradicting. This Section describes the main methods used in determining the equilibrium exchange rate and what the results from different papers are. Also it looks into the limitations of each theory. Finally this Section shows how this paper is similar to other papers and how it differs from previous studies. Most of the methods used by economists can be categorized into one of the following methodologies:
5.1 The Purchasing Power Parity Approach (PPP)
History:
Purchasing power parity (PPP) is one the oldest and simplest models used for exchange rate determination. The theory was first introduced by Swedish economist Gustav Cassel in 1918. The basic idea of PPP is that based on the law of one price the exchange rate between two countries is equal to the ratio of the price levels. (Krugman, Obstfeld, et al, 2011).
Theory in practice:
The goal of the PPP theory is the estimation of an equation for the real exchange rate, by looking at past behavior of the real exchange rate. According to the theory, over time a country’s nominal exchange rate should converge to its predetermined PPP level. In reality, certain factors prevent the nominal exchange rate from reaching this level, at least in the short run. One very influential paper on PPP describing this problem has been by Rogoff (1996).
Important factors deviation PPP level:
According to Rogoff ,the failure of short-run PPP can at least for some part be explained by the stickiness of nominal prices in the short run, as described by Rudiger Dornbusch (1976) in the sticky-price monetarist model. A second important factor to explain deviations from PPP was the Ballasa-Samuelson effect by Balassa and Ballasa-Samuelson in 1964. When a developing country like China is opening up to international trade, productivity in the tradable goods sector increases and so do the wages in this sector. As a result of this, wages in non-tradable goods sector also rise while
productivity in this sector remains more or less the same, thereby increasing the prices of non-traded goods. Under the assumption that prices of traded goods do not rise due to world price levels, overall price levels in the developing country rise faster than in the industrialized country. This should lead to an appreciation of the real exchange rate in the developing country, known as the Ballasa-Samuelson effect (Dunaway, Li, 2005).
Previous studies:
A number of studies were conducted to determine if the RMB moved in accordance with the exchange rate as predicted by PPP. One paper to investigate the equilibrium exchange rate of the RMB based on Relative Purchasing Power is by Chang (2007). He finds that after controlling for structural changes in the period 1976-2001 the RMB is undervalued by 22% in 2001,using annual data from WDI. Cheung, Chinn and Fujii (2007) perform a large cross country research using absolute PPP. They use data from 160 countries during the period 1975-2004 with data from WDI and perform a pooled time series cross section (OLS) regression. They find that once sampling uncertainty is accounted for, no substantial undervaluation of the RMB is present. A third very important article by Coudert and Couharde (2007) uses panel data on 132 countries between 2000-2004. They find an undervaluation between 16 and 29% depending on the chosen sub sample of countries.
11
A final remark on PPP is by Isard (2007). His research shows strong support for PPP as a long-run measure, but with an important side mark. The amount of undervaluation depends strongly on the chosen price index. Although all models predict undervaluation, the results differ from 10 to 40% depending on CPI, GDP deflator or PPI as a price index. He emphasizes that it is very difficult to determine the appropriate index for a country. This point is further made by an influential paper of Xu (2013).
5.2 Macroeconomic Balance (MB)/ Fundamental equilibrium exchange rate (FEER)
History:
The Macroeconomic Balance Framework (MB) framework was first introduced by the ideas of Ragnar Nurske in 1945 and Loyd Metzler in 1951. This MB approach looks at the requirements to achieve simultaneous internal and external balance in an open economy (Isard, 2007).
FEER:
The basic MB framework has been further studied, and since the research by Williamson (1994) became a method for determining equilibrium exchange rates. This way of determining the equilibrium exchange rate has become known as the Fundamental Equilibrium Exchange Rate (FEER) method. Here the equilibrium exchange rate is defined as the real exchange rate or as the real effective exchange rate. (Clark & MacDonald, 1998).
Effective exchange rate:
Because the FEER analysis also uses effective exchange rates, the concept of effective exchange rate must be illustrated. The regular nominal and real exchange rate are determined as the value of one currency to another between two trading countries. In general however, countries do not have a single trading partner. With multiple trading partners, for a country’s nominal effective exchange rate the domestic currency is valued against a basket of currencies. The weights in this basket are based on the amount of trading of one country with another country as a percentage of total trade (Pilbeam, 2013). The real effective exchange rate (REER) is a nominal effective exchange rate index adjusted for relative movements in national price or cost indicators of the home country, selected countries, and the euro area10. Although this effective exchange rate gives better understanding how the
competiveness for the industry of a country changes over time, data on effective exchange rates is less available in common databases.
Theory in practice:
The basic idea of the FEER approach is to look at the internal and external balance of a country’s economy and determine what the exchange rate should be to achieve these balances simultaneously, this exchange rate is then set as the equilibrium exchange rate. As an external balance the current account is used and for the internal balance either net flow of private and official capital (CAP) or domestic savings – investment (S-I) .If for instance the current exchange rate needs to appreciate to achieve equilibrium, the currency is undervalued (Li, Dunaway, 2005).
Previous studies:
Li and Dunnaway (2005) mention that the MB approach has a lot of potential but that it is very complicated to perform compared to PPP. In order to determine the value of the CU, CAP and/or S-I a lot of information is required. Also choosing the right components for each balance takes a lot of skill and knowledge. Isard (2007) mentions that a concern for MB approach is that especially for fast growing countries like China, the structure and composition of the economy changes over time.
12
Goldstein (2004) uses a simplified FEER with the equation CU=CAP and find a 4% CU surplus (of GDP) between the external and internal balance. He argues that to close this 4% gap an appreciation of the RMB between 15 and 30 percent is required. A study by Bosworth (2004) using a FEER approach with the identity CU=S-I, argues that the MB framework does not support the claim for fundamental RMB undervaluation. Finally an influential paper by Coudert and Couharde (2007) uses a complicated econometric FEER approach, with S-I for the internal balance, finding an
undervaluation of the RMB between 23 and 44%.
5.3 Behavioral Equilibrium Exchange Rate (BEER) model
Background:
A third category of models is the concept of Behavioral Equilibrium Exchange Rate (BEER). While the BEER approach also aims at calculating either the real effective equilibrium exchange rate or the real equilibrium exchange rate, it differs from the FEER approach in the previous sector. The BEER approach calculates the equilibrium exchange rate by direct econometric analysis using an appropriate set of explanatory variables .‘’The actual real exchange rate is said to be in equilibrium in a behavioral sense when its movements reflect changes in the economic fundamentals that are found to be related to the actual real exchange rate in a well-defined statistical manner’’(Clark & MacDonald, 1998). Zhang (2001) mentions that BEER analysis is well suited for developing countries because large and complex models often use lots of data which is not always available in these countries, this is also the case for China. He further stresses that for estimation of the BEER model theoretical guidance is needed for choosing the appropriate set of explanatory variables.
Theory in practice:
Using this approach, the equilibrium exchange rate is estimated using multiple chosen variables. The long run relationship between these variables and the exchange rate is interpreted as the equilibrium exchange rate. Clark and MacDonald (1998) present such an general reduced form expression with the following equation:
Qt = β1 * Z1t + β2 * Z2t + τ * Tt + εt
Z1 = vector for long run economic fundamentals
Z2 = vector for medium term economic fundamentals
T= vector short run factors
β1 + β2 + τ= coefficients long, medium, short term vectors
εt = random disturbance term
There is however an important aspect to the BEER analysis. To determine the possible under/over valuation of the exchange rate, the estimated value of the exchange rate Qc has to be sustainable. Qc
can only be sustainable if the underlying economic variables used to predict it are also stable, unfortunately this may not be the case in the short run. Therefore one has to calculate the difference between the actual real exchange and the exchange rate predicted by the long and medium run values of the economic fundamentals.
In formula form: Mt = Qr - Qc = Qr - (β1 * Z1t + β2 * Z2t ),
Mt stands for total misalignment and Qr is the actual real exchange rate. Most of the articles using a
BEER approach indeed look at data for a longer period of time. Not only to enhance the predictive power of the model but also to deal with this specific problem.
Some sort of theoretical guidance is needed for choosing the right variables. From the ideas of Clark and MacDonald, Zhang (2001) uses a BEER approach for the RMB, using four variables: gross
13
fixed capital formation as percentage of GDP, government consumption, growth rate of exports and the sum of exports and imports as percentage of GDP. He finds that before reforming the economic policy in 1981 the currency was overvalued. But that after 1981 the currency was undervalued most of the time, suggesting that the PBC used the exchange rate as an active policy tool. Another article using a BEER approach is by Bénassy, Duran and Lahrèche (2004) , they find an undervaluation of the RMB between 44 and 47 % for the period 2001-2003.
These previous Sections give a summary of the three main methods used for determining the equilibrium exchange rate. Although the PPP method has a proven empirical relationship in the long run, many papers find the method is too simplistic. Also the outcome of a possible currency
misalignment is highly dependent on the chosen price index, while justifying a certain index is not always possible. The FEER approach is based on achieving macroeconomic balance but requires a lot of data, which proves to be a problem for China. The method used in this paper therefore is a BEER approach based on a paper by Kim and Korhonen (2005). The next Section describes the model used in this paper, how it differs from current literature and how it aims to determine the possible
undervaluation of the RMB. Section 6: Data
The data on the 4 explanatory variables (gdpt, govt, capt, opent ) is annual data for the Chinese
economy, covering the period 1978-2012.The timeframe is chosen based on the availability of data. This data comes directly from the World Development Indicators supplied by the World Databank.
As a dependent variable this paper uses both the Real Exchange Rate (RER) and the Real Effective Exchange Rate (REER). Data on the REER is directly from the WDI in the form of an index (with 2005=100) for the period 1980-2012. Indirectly, the data of RER is acquired by calculation for the period 1978-2012, based on a slightly different definition than the one used in the literature review. Here we define the nominal exchange rate E as the units of domestic currency per unit of foreign currency, also Kim and Korhonen use this definition the other way around. RER is determined using the formula: RER= . Where P* is the price level in the US, E is the nominal exchange rate in units of RMB per U.S. dollar, P is the price level in China. Under this definition a decline in the real exchange rate denotes depreciation, while an increase in the RER indicates appreciation. Data on the nominal exchange rate also comes from the WDI.
For the price level in China and the US this paper uses a GDP deflator11 price index with 2010=100. The GDP deflator is non-seasonally adjusted and taken from the Oxford economics set in Datastream. The original paper by Kim and Korhonen does not specify which price index they use for the domestic and foreign price level. This paper uses GDP deflator because it accounts for changes in government consumption, capital formation ,international trade and the main component household consumption. A lot of these components are reflected in the chosen variables. Although CPI as a price index is widely used, this article chooses not to use CPI as a common price index. The main reason is that its composition can vary a lot between countries and is therefore useful for measuring inflation within a country, but of less value when comparing countries.
14 Section 7: Method
The goal of this Section is to estimate the possible undervaluation of the RMB for the period 1978-2012 using the RER and for the period 1980-1978-2012 using the REER.
7.1 Model
The model itself consists of four variables to estimate the long run relationship between the exchange rate and the behavioral fundamentals of the economy. The estimated model is:
REER,RERt = α + β1gdpt + β2govt + β3capt + β4opent
RERt , is the real exchange rate of the RMB for year t, as defined in Section 3.REER is the real
effective exchange rate of the RMB for year t. gdpt , is GDP per capita in current U.S. dollar for China
in year t. govt , is government consumption as percentage of GDP in year t. capt , is gross fixed capital
formation as percentage of GDP in year t. opent , is the sum of exports and imports as percentage of
GDP in year t.
The reasoning behind choosing these four variables to determine the long run equilibrium real exchange rate is based on the ideas by Hinkle and Montiel (1999). First, domestic supply-side factors should be included, with in particular variables relating to the Balassa-Samuelson effect. GDP per capita (gdpt ) serves as a measure for the Balassa-Samuelson effect. Gross fixed capital formation as
percentage of GDP (capt) also determines domestic supply capacity and is further used as a measure
for technological progress and foreign investment. Second, fiscal policy measures may affect the long run equilibrium exchange rate. Government consumption as percentage of GDP (govt) is used as a
proxy for fiscal policy changes. Third, other factors to include relate to changes in the international economic environment. Openness of the economy measured by the sum of exports and imports as percentage of GDP (opent), is included to measure the impact of commercial policy and the trade
regime.
Originally Kim and Korhonen (2005) use this model to estimate the real equilibrium exchange rate for a number of transition countries. With this model they test if these transition countries fulfill the criterion on exchange rate stability, in order to join the EU. They use an out of sample estimation using panel data on 29 middle and high-income countries from 1975-1999. With a Pooled Mean Group (PMG) estimator they estimate the value of each of the 4 coefficients and determine whether the exchange rate of the transition countries is at equilibrium level.
This article however does not perform an out of sample estimation but only looks at data on these 4 variables for China. For the transition countries in the original paper it is relatively easy to find countries that are similar to each other from an economical perspective. Most of the countries have similar industries, trade with the same countries and experience economic growth of about the same size. Because the coefficients are estimated using a BEER approach, using an out of sample
estimation makes sense. As explained in the previous Section, the goal of the BEER approach is to calculate the equilibrium exchange rate using variables that explain the behavior of an economy. If the economies ‘behave’ in the same way an out of sample estimation is useful. Unfortunately this is not so easy for China. Because of the economic developments in China over the past two decades finding a country that is similar to China from an economical perspective proves very difficult.
15 7.2 Pre-regression tests
Unit roots:
Because we are dealing with time series data we first have to test for unit roots. The unit root test I perform is the DF-GLS unit root test. The DF-GLS test performs a modified Dickey–Fuller t test for a unit root in which the series has been transformed by a generalized least-squares (GLS) regression. This test looks if it can reject the null hypothesis that unit root is present. If a unit root is present the data is said to be non-stationary, which causes the mean and variance to change over time. Although even with unit roots present least squares estimation is still consistent, shocks to a unit root process have permanent effects. The consequence of this is that the variance of a unit root diverges to infinity, influencing the standard errors and reducing the predictive power of the model. The results of the DF-GLS test for RER, REER, gdpt, govt, capt, opent are in Table 9. Unfortunately for almost all the
variables the presence of unit roots cannot be rejected, this can only be done for gdp at 10% significance level and open at the 5% level ,both with 3 lags.
VECM:
Standard regression techniques like OLS require that variables are covariance stationary. As shown by the unit root tests, most of the variables from this regression contain unit roots. The estimation and interpretation of non-stationary variables is possible with cointegration analysis. This analysis is based on the fact that time series variables that are not covariance stationary can be still be first difference stationary. These first differences can be modeled with a vector error correction model (VECM). Error correction models in general estimate the speed at which an dependent variable returns to its
equilibrium value after a change in one or multiple independent variables. To test for cointergration and use the VECM model in Stata, first the appropriate number of lags used must be chosen. The results in Table 10 show the appropriate amount of lags is 6 .
With the appropriate amount of lags known, the cointegrating rank of the VECM can be estimated using the Johansen test for cointegration. With 6 lags the null hypothesis of 0 cointegrating vectors can be strongly rejected but fails to reject the null hypothesis of at most 3 cointegrating vectors, see Table 11. When performing the VECM, Stata reports that there are too many parameters in the model, which means that there are not enough observations to deal with the degrees of freedom. Although we cannot use the VECM model, because there exists at least one cointegrating vector the model can be estimated using an OLS-regression.
7.3 OLS pre-estimation tests
The results of the VECM give a clear result, there are not enough data points to handle the method. Therefore this article continues by using an ordinary least squares (OLS) method, keeping in mind the presence of unit roots and the impact they have on the predictive power of the coefficients. Before running the OLS regression we test the residuals for homo/ heteroskedasticity , normality and autocorrelation.
The residuals are tested for normality using the Smirnov-Kolmogorov test. This tests the null hypothesis that the cumulative distribution of the residuals and the theoretical normal distribution are the same, using a chi-square test. Table 12 shows that a p-value of 0,1299 is not enough to reject the null hypothesis, so the residuals are normally distributed. Although this normality is not required in order to obtain unbiased estimates of the regression coefficient, it does assure that the p-values for the t-tests and F-test will be valid.
Next, to test if the residuals are homoskedastic the Cook-Weisberg is used. One of the main assumptions for the ordinary least squares regression is the homogeneity of the residuals. If the variance of the residuals is non-homogeneous the residuals are heteroskedastic and the OLS
16
regression with robust standard errors is used. This method looks if it can reject the null hypothesis that the residuals are homoskedatic. The outcome in Table 13 shows that a p-value of 0,0520 is not enough to reject the null hypothesis, so the residuals are homoskedastic at a 95% confidence level.
Last, to test for autocorrelation of the residuals the Ljung-Box Q statistic for autocorrelation of residuals has been used. The p-values are very low indicating the null hypothesis of no
autocorrelation is rejected, see Table 14. Serial correlation will not affect the unbiasedness or the consistency of OLS estimators, but it does affect efficiency. For instance, with positive serial
correlation, the OLS estimates of the standard errors will be smaller than the true standard errors. This will lead to the conclusion that the parameter estimates seem more precise than they really are
Wooldride (2012). Having done these tests, we can perform the OLS regression to estimate the coefficients.
Section 8: Hypothesis
Before looking at the regression results, this Section describes the expected outcomes for the sign and size of the estimated coefficients of both the RER and the REER. This prediction is based on the formula used to calculate RER and REER combined with the ideas of previous articles.
8.1 predicted coefficients RER
As described in Section 6, this paper determines the RER using the formula: RER= . Where P* is the price level in the US, E is the nominal exchange rate in units of RMB per U.S. dollar, P is the price level in China. Under this definition a decline in the real exchange rate indicates depreciation, while an increase in the RER indicates appreciation.
gdpt:
First we look at gdpt which serves as a measure for the Balassa-Samuelson effect. The literature
review shows that due to this effect, overall price levels in the developing country rise faster than in the industrialized country, leading to an appreciation of the real exchange rate in the developing country . This causal relationship between the Balassa-Samuelson effect and the RER uses the definition for the RER by Kim & Korhonen. According to their definition, an increase in domestic prices P, means that RER goes down. At the same time a decrease of RER could also be achieved by an increase of the nominal exchange rate E. They value E as units of foreign currency per unit of domestic currency, so if E goes up more foreign currency can be obtained with one unit of domestic currency and the currency appreciates. However for the RER definition we use, the nominal exchange rate E is formulated the other way around. A RMB appreciation means that less units of the RMB are needed for one dollar, so E decreases. With a RMB appreciation ,E goes down and the RER index goes up. According to our definition, if the domestic price level increases RER goes up. We still use the idea behind the Ballasa-Samuelson effect as described by Kim & Korhonen (2005), only instead of having a negative coefficient for gdp, this article expects to find a positive coefficient. To
summarize, with an increase in GDP per capita domestic price level goes up leading to appreciation. Under our definition appreciation means that E goes down and causes the RER to go up. Therefore I expect a positive sign for the variable gdpt.
capt:
The variable capt is used to measure technological progress and foreign investment. If there is more
technological progress within a country, we expect overall price levels in that country to decrease. New technologies for instance, could lead to greater cost efficiency, therefore reducing prices. An increase in foreign investment would practically do the same. With increased investment the amount of capital per worker increases, resulting in higher output and possibly lower costs. So if foreign
17
investment and technological progress go up, increased cost efficiency in China means that its relative price level compared to the U.S. decreases. For the variable gdp we showed that with an increase in the domestic price level, RER goes up. Here we see that domestic price level decreases so RER goes down. As an increase in gross capital formation causes the RER to decrease, I expect a negative sign for the variable capt
govt:
This variable is used as a proxy for government spending to model the effect of fiscal policy. Most of the government expenditure is service orientated, labor intensive and stays within the country. For example, expenditure on healthcare and infrastructure stays within the country. Due to the high labor insensitivity, increased government spending increases the demand for labor. This increased demand for labor means that wages in this sector are being pushed up. As a result, wages in other sectors also increase. Producers increase their prices and the domestic price level slowly goes up causing RER to go up. As an increase in government spending causes the RER to increase, I expect a positive sign for the variable govt.
opent:
Last, we look at opent, which is included to measure the impact of commercial policy and the trade
regime. A closed economy is often associated with overvaluation. Therefore, If the country becomes more open to foreign trade, ceteris paribus, the domestic currency should depreciate to remain competitive. Under our definition, depreciation means that E goes down and the RER goes up. The more open the economy of a country, the more the its currency should depreciate , increasing the RER. Therefore this article expects to see a negative sign for the variable opent.
8.2 predicted coefficients REER
Using the REER instead of the RER as an dependent variable in the regression, the signs of the estimated coefficients are expected to remain the same, although the size could differ. The causal relationship these variables have should not be much different for either RER or REER. The only difference between them is that the REER is based on the relative amount of trading with each trading partner.
When determining the RER we only looked at the price levels of China and the US. For the REER we should look at all China’s trading partners. For example, in 2012 the main trading partners for China were: EU, USA, Hong Kong, Japan and South Korea. Just these 5 trading partners
accounted for over 50% of China’s total trade12
. The EU, Hong Kong, Japan and South Korea are all major industrialized countries and are quite similar to the US from an economical perspective. Also these 5 trading partners alone account for the majority of China’s total trade. We expect not much difference between the estimated coefficients for the RER and the REER because the other trading partners needed to compute the REER are quite similar to the US and make up for the largest part of China’s total trade. Looking at 2012 gives a snapshot, the idea behind it however can be used for previous years. Although China’s trading partners have changed over time most of them were still large industrialized countries and therefore comparable with the US.
12 Source: Eurostat IMF
18 Section 9: Results
9.1 Results RER
Table 15. OLS estimates RER, 1978-2012:
Table 7 shows the results of the OLS regression with RER as a dependent variable. Unfortunately only the variable opent is significant. It seems the data does not show anything because of constraints
being put on the model due to a lack of data. Still we are able to take a look at the signs and the size of the coefficients.
The variable opent shows a negative coefficient that is consistent with theoretical predictions.
This could mean that also for China, opening up its economy to international trade resulted in a depreciation of its currency to remain competitive. In this case a 1% increase in the sum of exports and imports as percentage of GDP , resulted in a depreciation of 1,44 % of the real exchange rate. The fact that opening up to international trade caused depreciation in China is described by Rodrik (2010). He mentions that a big change took place when the country joined the WTO in 2001. Before joining the WTO its economy implemented lots of protective measures. By joining the WTO this was no longer possible and China had to look for other ways or remaining competitive and realize economic growth. Zhang (2001) and Kim & Korhonen (2005) find the same sign for the coefficient opent with
the values -0,87 and -0,65.
Interestingly enough govt does not behave according to the theory. The negative sign indicates
that a 1% increase caused a 0,844% depreciation of the RMB. On explanation would be that the increased demand for labor did not increase wages. Another explanation could be that to remain competitive ,producers decided not to increase prices but tried to cut additional costs or reduce profit margins. Both Zhang (2001) and Kim & Korhonen (2005) find that an increase in government spending leads to currency appreciation. Apart from the coefficient being very insignificant, at this time we cannot give a definitive answer for this.
On the other hand capt moves in line with the predictions. According to its coefficient a 1%
increase in gross capital formation as percentage of GDP depreciated the real exchange rate by 1,38%. The coefficient shows that it is possible for increases in technological progress and foreign investment as measured by cap, to have increased cost efficiency. Interesting is that while Zhang (2001) shares this result, Kim & Korhonen find that capt has an opposite sign.
Last, looking at the variable gdpt it has the sign predicted by the Balassa-Samuelson effect in
the literature review. When a developing country like China is opening up to international trade, prices of non-tradable goods rise relatively faster compared with traded goods, resulting in a RMB appreciation. The coefficient for gdpt is small due to a matter of scaling. All other variables (govt,
19 9.2 REER
Table 16. OLS estimates REER, 1980-2012:
For the REER we experience the same problems as with the RER regression. Unfortunately due to the lack of available data and possibly the existence of unit roots and serial correlation only the variable opent is significant.
The variables open, cap and gdp have the same signs as the regression with RER and thus those predicted by theory, however the size of the coefficients is different. First, open depreciates the equilibrium exchange rate less when estimations are made with REER. Second, the coefficient of cap depreciates the equilibrium exchange rate more. The findings for both gdp and cap are shared by Kim & Korhonen (2005). Surprisingly the effect of the Ballasa-Samuelson effect as measured by gdp is bigger in this paper when using REER, Kim & Korhonen find an opposite result.
What stands out most in these OLS-estimates is the sign of govt. Before performing the
regression we expected to only see changes in the size of the coefficients. Two changes occurred for the variable gov when using the REER. One, the variable now moves in accordance with theory, as a rise in government spending resulted in an appreciation of the RMB. Two, when using the REER the impact of the variable gov became bigger which is also seen in the results of Kim & Korhonen. One possible explanation for the sign change could be that Chinese producers could not raise prices when trading with the US, but that they were able to do this with other major trading partners. This because prices of Chinese exports were already high in the US due to protectionist measures and less of these measures existed in other countries. However, increased globalization makes the existence of different prices for the same product unlikely. Further research should be done to give a definitive answer to this.
9.3 misalignment of the RMB, RER
This paper started by asking the question whether the RMB was undervalued. The coefficients for both the RER and REER have been estimated using OLS. Although most of these coefficients were not significant we still estimate the undervaluation and compare this with the developments in the exchange rate regime described in Section 2. Based on the OLS estimates we get the following model for RER:
RERt = 1,465 + 0,0000391gdpt – 0,844govt – 1,385capt – 1,4465opent
The results are shown in Table 17. Looking at these there are a number of interesting observations to be made. According to this model the RMB was overvalued until 1984. Starting in 1985 it became
20
undervalued, this is also the beginning of a period in which several large devaluations of the RMB compared to the dollar took place. Second, the results show the currency became even more
undervalued in the period 1994-2004. In 1994 China indeed further reformed its foreign exchange rate system which was accompanied by a large depreciation. The RMB remained undervalued between 1994 and 2004, defined by the PBC as a period of managed float. Third, the results show an appreciation of the RMB since 2005. Suggesting that pegging the RMB to a basket of currencies brought it closer to its equilibrium value. Estimating the undervaluation of the RMB using the RER we see that the undervaluation seems to follow the changes in the Chinese exchange rate regime during the period 1978-2012. However, large deviations from the general trend still exist. In both 2007 and 2010 the results show a sudden overvaluation and although the RMB seemed to be appreciating since 2009, the Chinese currency was again greatly undervalued in 2012. Based on the results from the RER this paper supports the claims by foreign officials that the RMB has been below its equilibrium level for the period 1985-2012, except for 2007 and 2010.
9.4 misalignment of the RMB, REER
Based on the OLS estimates we get the following model for REER: REERt = 1,235 + 0,0000761gdpt + 1,484govt – 1,519capt – 1,0398opent
The results are shown in Table 17 and show a number of interesting developments for the RMB. The RMB was overvalued until 1985 and undervalued in the period 1986-2002. This result is quite similar to the estimated under/over valuation using the RER, with two exceptions. One, in almost every year estimated undervaluation of the RMB using the REER was less that the undervaluation using the RER. Two, undervaluation did not go up in 1994, in fact it became less. Another striking feature is that with the REER the RMB was estimated to be overvalued for the period 2003-2010, with the overvaluation being roughly 70% for the years 2004,2005 and 2006. This is the exact opposite result from the RER, which shows an constant undervaluation for the same period. A possible reason could be the different sign for the estimated coefficient govt. However according to the data, government
spending did not change enough for the period 2003-2010 to explain the difference. Further research is needed to fully address this problem. Finally, according to the REER the RMB was undervalued in 2011 and 2012. So based on the REER this paper supports the accusations of an undervalued RMB for the period 1986-2002 and 2011-2012. That being said, for the periods 1980-1985 and 2003-2010 the methods shows the RMB was in fact overvalued.
Section 10: Conclusion and Discussion
10.1 Discussion
As may be expected, the results from this regression are not optimal. Because this article uses time series data, tests for unit roots had to be conducted. The null hypothesis of unit roots was not rejected making the data non-stationary. Shocks to a unit root process have permanent effects, making the variance of a unit root diverge to infinity and reducing the predictive power of our model. The data did show to be stationary at a first difference level, making it possible to use the VECM. When performing the VECM, Stata reported not enough observations were available to deal with the degrees of freedom in the model. Although using the VECM model was not possible, the existence of at least one cointegrating vector meant the model could be estimated using an OLS regression. Basic pre-regression tests showed the residuals were normally distributed, homoskedastic and serial correlated. Serial correlation will not affect the unbiasedness or the consistency of OLS estimators but it does affect efficiency, making the parameter estimates seem more precise than they really are.
21
Apart from unit roots and serial correlation, sample selection bias may also be regarded as a possible cause of insignificant coefficients. This sample selection bias comes from each variable’s small observation size of only 35 for the RER and 33 for the REER. One way to improve this significance is by using quarterly instead of yearly data, however this data was unavailable for China. Another way is performing an out of sample estimation for the coefficients as done by Kim & Korhonen (2005). Before doing this, considerable research has to be done in finding economies that behave in a similar way compared to the Chinese economy. Only then does an out of sample estimation using a BEER approach acts a valid extension.
The other possible explanation for insignificant coefficients is omitted variable bias. The Chinese economy like any other is very complex. In the regression model only 4 variables are taken into account to model the behavior of the exchange rate in the long run. Thus another possibility for future research could be improving the significance of the variables by introducing more and/or different variables to model the behavior of the economy.
Overall there is still a lot of discussion within current literature about the true definition of the equilibrium exchange rate and how it should be determined. Considerable work lies ahead in understanding the interaction between the various methods to obtain equilibrium exchange rates (Kim and Korhonen, 2005).
10.2 Conclusion
The aim of this paper has been to actively contribute to the discussion on the Chinese exchange rate, using a regression model based on a BEER approach by Kim and Korhonen (2005). The RMB has been the subject of controversy for a long period of time and is currently under pressure of
revaluation. In particular Japan and the United States fear that the long term undervaluation has given the Chinese economy an unfair comparative advantage in international trade.
China has seen some large changes in its exchange rate policy. The most important being opening up to international trade in 1978, unification into a single exchange market in 1994 and pegging the Renminbi to a basket of currencies in 2005. Two main arguments have been used by foreign officials to point out a undervalued RMB. One is the presence of large simultaneous surpluses on both on the current and the financial account. The second is the rapid increase of foreign exchange reserves by the PBC. For both these arguments we stress that conclusions cannot be made directly and alternative explanations could be given instead of an undervalued RMB.
Next, this article has determined the effects of a possible RMB misalignment for the Global and Chinese economy. The main concern for the Chinese economy is that a possible future
appreciation would decrease the value of its foreign reserves, which accounted for 40,25 % of GDP in 2012. A major concern for the global economy is that a RMB deviating from its equilibrium level could cause current account imbalances in other countries, potentially triggering a new financial crisis.
Methods used by economist to determine the equilibrium exchange rate are categorized into 3 general methodologies. Research using the PPP method finds evidence for undervaluation of the RMB between 16 and 29%. However no appropriate guidelines exist for determining the right price index for a certain country. Papers using the FEER approach estimate the undervaluation of the RMB to be between 15 and 44%. Choosing the right components for each balance takes a lot of skill and knowledge, which is even more difficult for developing countries like China. The method used in this article is part of the third category called a BEER approach. Theoretical guidance is needed for choosing the right variables. Based on the the ideas of Zhang (2001) and Kim & Korhonen (2005) this paper uses 4 variables to implement a BEER approach.
22
The model to estimate the real exchange rate and the real equilibrium exchange rate is explained by GDP per capita, government consumption as % of GDP, sum of exports and imports as % of GDP and gross fixed capital formation as % of GDP. This model was estimated using an OLS regression. The results however show that for both RER and REER only the variable open proved to be significant.
However, almost all estimated coefficients did show the sign and size expected by the theory. Using these variables to predict the RER only the variable govt did not behave as expected. The
negative sign indicated that an increase in government spending caused an unexpected depreciation of the real exchange rate.
In an attempt to answer the research question this paper used the estimated coefficients for the RER and the REER to estimate whether the RMB was undervalued. A BEER approach using the RER, estimates the RMB has been overvalued between 1978 and 1984 and mostly undervalued for the period 1985-2012. The BEER approach using the REER finds a contradicting result for the period 2003-2010 in which the RMB according to this method was overvalued.
23 Reference list:
Amiti, M., & Freund, C. (2010). The anatomy of China's export growth. In China's growing role in
world trade (pp. 35-56). University of Chicago Press.
Baffes, J., Elbadawi, I. A., & O'connell, S. A. (1997). Single-equation estimation of the equilibrium
real exchange rate (No. 1800). World Bank, Development Research Group.
Bénassy-Quéré, A., Duran-Vigneron, P., Lahrèche-Révil, A., & Mignon, V. (2004). Burden sharing and exchange rate misalignments within the Group of Twenty. Dollar Adjustment: How Far?
Against What.
Bosworth, B. (2004, February). Valuing the renminbi. In Tokyo Club Research Meeting,
February (pp. 9-10).
Chamon, M. D., & Prasad, E. S. (2010). Why are saving rates of urban households in China rising?. American Economic Journal: Macroeconomics, 93-130.
Chang, G. H., & Shao, Q. (2004). How much is the Chinese currency undervalued? A quantitative estimation. China Economic Review, 15(3), 366-371.
Chang, G. H. (2007). Is the Chinese Currency Undervalued? Empirical Evidence and Policy Implications. International Journal of Public Administration, 30(2), 137.
Cheung, Y. W., Chinn, M. D., & Fujii, E. (2007). The overvaluation of renminbi
undervaluation. Journal of International Money and Finance, Vol.26, No.5, pp.762-785
Chou, W. L., & Shih, Y. C. (1998). The equilibrium exchange rate of the Chinese Renminbi.
Journal of Comparative Economics, Vol.26, No.1, pp.165-174
Clarke, P., & MacDonald, R. Exchange Rates and Economic Fundamentals: A Methodological Comparison of BEERs and FEERs. IMF Working Paper Series.
Coudert, V., & Couharde, C. (2007). Real equilibrium exchange rate in China is the Renminbi undervalued?. Journal of Asian Economics, 18(4), 568-594.
Corden, W.M. (1994). Economic policy, exchange rates, and the international system. Oxford university press.
Frankel, J. (2006). The Balassa-Samuelson Relationship and the Renminbi
Harvard WP.
Frankel, J. A., & Wei, S. J. (2007). Assessing China's exchange rate regime.
Economic Policy, 22(51), 575-627.
Gartner, M. (2009), Macroeconomics, 3th edition. Prentice Hall. Goldstein, M. (2004). Adjusting China's exchange rate policies.
24
Goldstein, M., & Lardy, N. (2006). China's exchange rate policy dilemma.
The American economic review, pp. 422-426
Hinkle, L. E., & Monteil, P. J. (1999). Exchange rate misalignment: Concepts and measurement for
developing countries. Oxford University Press.
Husted, S., & MacDonald, R. (1998). Monetary-based models of the exchange rate: a panel
perspective. Journal of International Financial Markets, Institutions and Money, 8(1), 1-19. Isard, P. (2007). Equilibrium Exchange Rates: Assessment Methodologies.
IMF Working Papers, 1-48.
Kim, B. Y., & Korhonen, I. (2005). Equilibrium exchange rates in transition countries: Evidence from dynamic heterogeneous panel models. Economic Systems, 29(2), 144-162.
Krugman, P., & Obstfeld, M. Melitz Marc (2011). International Economics: Theory and Policy, 9th
edition. Prentice Hall.
Lee, J., Bayoumi, T., & Faruqee, H. (2005). A fair exchange? theory and practice of calculating
equilibrium exchange rates (No. 5-229). International Monetary Fund.
Li, X., & Dunaway, S. V. (2005). Estimating China's" Equilibrium" Real Exchange Rate (No. 05/202). International Monetary Fund.
Ma, G., & Yi, W. (2010). China's high saving rate: myth and reality.
International Economics, 122, 5-39.
Max Corden, W. (2009). China's Exchange Rate Policy, Its Current Account Surplus and the Global Imbalances*. The Economic Journal, Vol.119, No.541, pp.430-441
Rodrik, D. (2010). Making room for China in the world economy.
The American Economic Review, 89-93.
Rogoff, K. (1996). The purchasing power parity puzzle.
Journal of Economic literature, 647-668.
Wang, T. (2005). Sources of real exchange rate fluctuations in China.
Journal of Comparative Economics, Vol.33, No.4, pp.753-771
Wooldridge, J. (2012). Introductory econometrics: A modern approach. Cengage Learning. Williamson, J. (Ed.). (1994). Estimating equilibrium exchange rates. Peterson Institute. Zhang, Z. (2001). Real exchange rate misalignment in China: An empirical investigation.
25
Appendix, Figures and Tables:
Figure 1. China’s nominal exchange rate, RMB to the U.S. dollar, 1985-2012
Source: People’s Bank of China, http://www.stats.gov.cn/tjsj/ndsj/2013/indexeh.htm
Figure 2. China’s nominal exchange rate, RMB to the U.S. dollar, 2005-2014
Source: Yahoo Finance, http://finance.yahoo.com/echarts?s=usdcny=X
0 1 2 3 4 5 6 7 8 9 10 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
