Euro and the yuan: different peas in the same pod.
Two of the important concerns in world economies and financial markets are the value of the yuan and the value of the euro. The latter is not a problem currency for the European Community at large but it is for the Tier II economies of Greece, Ireland, Italy, Portugal, and Spain (hereafter, the euro Tier II economies). There are obvious differences between the two global concerns. The China problem is a problem of plenty--there is too much savings, the growth is too strong and there is likelihood of overheating. The euro Tier II situation is one of too much consumption, too little growth, of large current account and government deficits, and of a large national debt. The two problem currencies may seem dissimilar at the surface, but there is a strong likelihood of both being associated with the 'same' policy, and therefore involve the same policy 'solution'. Both involve misalignment of the currency--the yuan is undervalued for the world, and the euro is too overvalued for the Tier II countries. With the yuan undervalued, there may be high growth and current account surpluses; with the euro overvalued for the Tier II economies, there is a likelihood of slower growth and large current account deficits.
The Tier II economies have to suffer or enjoy the same exchange rate as the rest of the members of the euro area. Unfortunately, despite there being both a political and economic union, the confederation of European states is not a federation. In the latter case, large countries like USA, China and India can and do compensate the poorer states for the 'losses' that emanate from very disparate per capita incomes and different levels of competitiveness under the umbrella of the same exchange rate. The resolution of the Tier II debt crisis (or an exchange rate problem in disguise) is likely to involve some version of this transfer compensation. It is reasonable to assume that had these countries not been tied to the rich country euro, the problem would have been much less severe, and most likely not have emerged into a full blown crisis.
This paper is an attempt to distill and integrate findings on the role that exchange rates, and their valuation, play in determining the fortunes and deficits of nations. In Bhalla (forthcoming) I present historical and current evidence to support the proposition that exchange rate undervaluation has played an important role in development and growth, and that its 'management' has been an explicit goal of policy making for several countries since the late 19th century. Further, that valuation of the exchange rate affects current account balances (CABs). There are other determinants of current account imbalances, but perhaps none as important as the valuation of the currency--ceteris paribus, exchange rate undervaluation positively affects the CAB, and overvaluation negatively affects CAB. The euro and the yuan problems are one of imbalances--the undervalued nature of the yuan is associated with large current account surpluses in China, and the overvalued nature of the euro, at least for the Tier II economies, may be associated with current account deficits in these economies.
The plan of the paper is as follows. The next section summarizes the theoretical and empirical issues associated with the measurement of currency valuation, and the relationship of valuation to growth and imbalances. The subsequent section examines and compares the experience of two major surplus economies--Japan in the early 1980s and China post 2001. The penultimate section provides an interpretation of the dual nature of current account surpluses in Tier I economies, and current account deficits in the Tier II economies. The final section concludes.
MEASUREMENT OF REAL EXCHANGE RATES
That the misalignment of important currencies is an important determinant of global economic problems is a tall conclusion, yet a defendable one. Of course other determinants matter but the simple point is that currency misalignment might be a bigger cause than recognized, and that its tentacles affect other more weighty indicators of financial crises like debt or the fiscal deficit.
This assertion puts the burden of proof on the measurement of currency misalignment. Only if currency misalignment is correctly measured, can one move to the second stage of the analysis--the effects of misalignment on growth and imbalances. How does one determine whether a currency is over- or undervalued? There are various indirect indicators, and none of them work all the time. The most obvious is to determine whether the exchange rate would lead to, in the long run, a current account surplus or deficit. In the surplus case, the exchange rate is most likely undervalued, that is the currency in question is so cheap that it leads to an excess of exports over imports. Indeed, the method of using the magnitude of the surplus/deficit to derive equilibrium exchange rates was pioneered by John Williamson (1994) in the mid-1990s and termed the 'Fundamental Exchange Rate Equilibrium' or FEER. While popular and with intuitive appeal, this method relies rather heavily on the target of close to zero balance in the current account. Countries can have excess savings, and therefore a positive CAB for a variety of reasons, for example Germany, Japan and China. Expecting exchange rate movements to solve such imbalances may be unrealistic. Nevertheless, the FEER method is broadly correct in identifying under- or overvaluation as an important determinant of current account imbalances.
There is an alternative Balassa method of estimating equilibrium exchange rates. This method relies on a stylized relationship between the real exchange rate (RER) and per capita income. As a country becomes richer, its RER increases. This observation was made, in separate papers, by both Balassa (1964) and Samuelson (1964). Both were trying to explain why the US dollar was not overvalued in the 1960s. (1) Both explained how the overall price level tended to increase with development, primarily because of a lower than average increase in productivity in the services sector. This meant that the RER was higher in richer countries, where the RER is defined as the ratio of price levels with the numeraire economy, the USA.
These two methods can yield estimates for equilibrium exchange rates and deviations thereof. The Balassa method is relatively easy to implement, and has been used by several authors, for example Dollar (1997), Easterly (2005), Bergin et al. (2004), Bhalla (2006, 2008), Johnson-Ostry-Subramaniam (2007), hereafter JOS, and Rodrik (2008). Implementation of the Balassa method involves the estimation of the relationship between RER, measured as the ratio of the PPP and US dollar exchange rates, and per capita income (measured in 1996 PPP dollars). (2) The predicted value of RER from this relationship is assumed to be the 'equilibrium' exchange rate. Since the data used are across-time and across-countries values, what the predicted value communicates is a stylized relationship between average RER and average per capita income.
In equation form, the estimated relationship is
RER = f(Y) (1)
where RER is the real exchange rate and Y is real per capita per day income in 1996 PPP dollars. The equation yields RER* as the predicted 'equilibrium' value. Currency over- or undervaluation (UV) is defined as the (log) deviation of RER from its predicted value:
UV = 100 x log(RER/RER*) (2)
If actual RER is higher than equilibrium RER*, then the currency is overvalued, and UV is positive; if less, the currency is undervalued and UV has a negative sign.
Convergence and divergence in functional forms relating RER to income
Theory does not suggest the functional form for equation 1. Balassa assumed the relationship to be linear; Dollar and Easterly assume it to be semi-linear (income is expressed as income and income squared) while most others assume the relationship to be log-linear.
In Bhalla (2006, 2007, forthcoming) I attempt to derive a functional form from stylized facts about economic growth. Part of the old conventional wisdom (and still applicable today) was that income levels followed a nonlinear S-shaped path. This was primarily due to the reallocation of labor from agriculture to industry. In the early stages of development, when the share of agriculture is upwards of 60%, this movement from a sector growing at say 3% to a sector growing at perhaps 6% does not much affect overall GDP growth. Hence, in these early years, per capita income evolves in a flattish fashion--one end of the S. Once the agriculture/GDP share drops to less than 50%, there is an acceleration in overall growth due to the reallocation of labor--this is the steep part of the S. And when a country becomes developed and/or the share of agriculture drops to less than 25%, the reallocation effect is about over--this is the upper flat end of the S.
If one looks at various examples, and explanations, of growth, it is remarkable how so many of them trace out the S-shape. (3) As early as the 1950s, Rostow (1960) offered his stages of growth; the third stage is the takeoff stage, which corresponds to the steep slope area of the S. The final stage, a mature economy, is the flat portion of the S.
Bhalla (forthcoming) discusses reasons, and provides evidence, that the relationship between the RER and per capita income is also likely to be S-shaped. The S-shaped pattern is suggestive of what is termed, in jargon, as an 'exponential form with one asymptote'. It is represented by the functional form RER = [b.sub.1](1-[b.sup.Y.sub.2]), where b1 and b2 are parameters to be estimated. By construction, the S shape converges to [b.sub.1] when per capita income Y becomes very large (as long as b2 is less than 1).
Estimation of this S-shaped functional form results in an exceptionally large explanatory power R2 equal to 0.87. The estimated non-linear relationship for the S-shaped curve for 184 countries for the time period 1996-2008 (annual data) is as follows (4):
RER = 1.12(1 - [0.971.sup.Y]), Nobs = 2321, [R.sup.2] = 0.87 (3)
The average convergence level of RER is 1.12, that is as incomes go beyond a certain level, the equilibrium real exchange rate stays approximately constant (see Figure 1).
An alternative specification, a log-log model, forces a constant relationship between RER and per capita income. This specification (estimated by JOG and Rodrik, among several others) yields the following estimate for the same set of observations:
Log(RER) = -1.83 + 0.37*log(Y), Nobs = 2321, [R.sup.2] = 0.47 (4)
[FIGURE 1 OMITTED]
The coefficient on log Y, 0.37, is constant across all values of per capita income. In the exponential model above, the elasticity varies with income--flat at first, then increasing and large, and then flat again. In the log-log model, the real exchange rate elasticity is the same (0.37) for Ethiopia (per capita income level of PPP$ 926, 1996 prices in 2008) and USA (per capita income level of PPP$ 36,680 in 2008). In the S-shaped model, the elasticities are 0.96 (Ethiopia) and 0.16 (USA).
By forcing the elasticities across different levels of income to be the same when the elasticities are likely to be different, the log-log model introduces measurement error. The constant elasticity yielded by the log-log model can be thought of as a weighted average of the 'true' elasticity along different levels of per capita income. A synthetic [R.sup.2] (calculated as the [R.sup.2] or variation in RER from equation 4) of 0.58 for the log-log model is considerably lower than the 0.87 obtained for the S-shaped curve. This large difference reflects the effect of the large specification error introduced into the model by the assumption of a constant elasticity between RER and income.
Figure 1 shows the relationship between RER and (log) per capita income for 2009; the same pattern is observed for different years over the last decade. The proportional distance away from the line measures the percent magnitude of the under- or overvaluation. Most of Europe has an overvalued currency along with most of sub-Saharan Africa--the actual values are considerably above the predicted line. Countries of East Asia are distinguished by being some distance below the predicted line--the currencies of these countries are undervalued. Sub-Saharan Africa has had problems of slow growth; East Asia has had problems of fast growth. The conclusion is suggestive and perhaps self-evident.
Effect of currency undervaluation on growth (and other variables)
Estimates of currency valuation as obtained from the above equations can be used to test various hypotheses relating such valuation to investments, CABs, and economic growth. However, as pointed out by Woodford (2008), the results pertaining to economic growth and currency valuation may have a 'construction' bias as the estimate of undervaluation is correlated with economic growth. In these models, for example Rodrik (2008), the predicted real exchange rate is obtained from a regression of the real exchange rate on income levels; and income levels in the present period are a function of the preceding period's growth rate. There are three ways of getting around this endogeneity--first, econometrically, via use of generalized method of moments (GMM) estimation; second, by using initial undervaluation as a determinant of future growth; and third, by using dependent variables that are not contaminated by the inherent feedback from last period's growth to present period's income, for example the share of investment in GDP as a dependent variable.
There are theoretical reasons to expect that currency valuation, however measured, affects investment decisions. The direct critical relationship is between currency undervaluation and investment. The commonly estimated relationship between undervaluation and economic growth is a reduced form relationship and should be viewed as such.
Bhalla (forthcoming) discusses how investment decisions are affected by both the level of currency undervaluation, and its change. But what determines investment? The obvious candidate is profitability, both absolute (the rate of return is higher than the cost of capital) and relative (the rate of return is higher than the return on the next-best alternative investment). Regardless, the profitability of the investment is the key, and currency undervaluation plays a key role in determining profitability.
Currency undervaluation directly affects the cost of labor, an important input in the production process. Today, the cost of capital is nearly the same for most foreign investment, though as Clark (2007) documents, this was also the case in the 19th century. So what differentiates the appeal of one country versus another for foreign direct investment? There are several features of an economy that make investment attractive: lack of red tape, quality of workers, tax regime, etc. However, a dominating influence in the decision is likely to be the cost and profitability of the investment relative to other countries. If the currency is overvalued, the costs of operating would be more than in another country, and profitability and foreign investment would be less, ceteris paribus. So currency undervaluation plays a direct role in the decisions about both who gets and who does not get foreign investment. The same cost and profitability logic applies to domestic investment.
Table 1 presents illustrative results for two dependent variables--average investment rate and per capita income growth. The sample consists of 115 countries for sequential 5-year periods starting in 1980. For each observation, the dependent variable (investment rate, income growth) is the average for the subsequent 4-year period and the independent variable, currency undervaluation, is from equation 3. For example, the data for the 5-year period 1980-1984 would have currency undervaluation and log initial per capita income for 1980 and income growth for the years 1981 to 1984. Use of initial undervaluation rather than the mean undervaluation in the period prevents 'construction' bias a la Woodford.
Results are reported for both a country fixed effects model and a GMM model. The latter econometrically controls for any endogeneity among the variables. The results are strong and unequivocal. For the investment share, the coefficient on initial undervaluation is about -0.04 (ie for each 10% increase in undervaluation, the share of investments in GDP goes up by almost one-half percentage point). The significance in both cases is at the 1% level. For per capita GDP growth, the coefficient is either -0.007 (fixed effects) or -0.014 (GMM). The results suggest an approximate 0.1% increase in average per capita GDP growth for each 10% of currency undervaluation--and a lower 0.1% average GDP growth for each 10% of overvaluation. (5)
RER and current account imbalances
CABs are presumed to be a function of the valuation of the currency--more undervaluation, ceteris paribus, leads to a higher CAB. (6) This simple relationship can be expressed as follows:
[X.sub.it] = a + b[UV.sub.it] (5)
where X is the CAB as a percentage of GDP, and UV is the level of undervaluation as derived from equation 4. The subscript it refers to the ith country for the tth time period. The relationship between the real exchange rate and the CAB is robust and stable. A simple regression between the two variables yields a coefficient close to minus 0.03 (Table 1), that is each 33% level of initial undervaluation, ceteris paribus, is equivalent to, or consistent with, a 1 percentage point increase in CAB as a share of GDP.
One additional result on currency valuation--devaluing while standing still
Popular talk, especially in the media and investment banks, start and stop discussion on currency misalignment with reference to the real effective exchange rate (REER), defined as the exchange rate adjusted for inflation rate differences between two countries or one country and a set of other countries. This narrow interpretation of the real exchange rate is facilitated by the easy availability of data on REER; international agencies like the International Monetary Fund (IMF) and the Bureau of International Investments (BIS) provide time-series data of REER for several countries. (7) Most often, the data on REER are against a basket of currencies rather than just the US dollar, and with trade shares as weights. The Indian central bank publishes, on a regular basis, its own REER measure for the rupee. Thus, the conventionally defined REER has official backing, and sanction.
In reality, there are two distinct components of valuation change--direct and indirect. The direct effect is the popularly measured deviation of a currency value from its REER. The choice of one or several countries in the computation of the REER country basket does not alter the fact that these computations (as done by IMF and BLS) completely miss out on the entire Balassa-Samuelson revolution. By giving zero weight to relative productivity changes, the REER measure fails to capture potentially the most important component of changing costs, and changing competitiveness--the indirect effect on valuation via differences in productivity growth.
If productivity growth was the same in all economies, then REER would be an appropriate relative cost measure. But productivity differentials are not the same, and indeed developing country productivity growth is likely to be higher. When 'excess' productivity growth occurs, the costs of production decline by an equal (percentage) amount. Such excess productivity growth, weighted by the elasticity of RER with respect to income, and when not accompanied by a revaluation in the currency, is a real devaluation. This is the indirect, or 'standing still', component of a real devaluation. The two components of valuation, direct and indirect, are additive in the logs. In summary, there are three components of real exchange rate change: nominal exchange rate change, relative inflation rates, and the contribution of relative productivity growth. The first two components are combined in the direct effect, [RER.sub.direct]; the third (indirect) effect is reported as standing still or [RER.sub.sstill].
In equation form, changes in currency valuation can be decomposed as follows (all changes are log changes denoted by the prefix d):
d[UV.sub.t] = [UV.sub.t] - [UV.sub.t-1] (6a)
dREER = [REER.sub.t] - [REER.sub.t-1] (6b)
d[UV.sub.direct] = dREER = d[XR.sub.t] + d[PGDP.sub.t] - d[PGDP.sub.t, usa] (6c)
d[UV.sub.standing still or sstill] + d[UV.sub.direct] = dUV (6d)
where XR is the exchange rate with respect to the USA and dXR is positive when the domestic currency appreciates, and PGDP and [PGDP.sub.usa] are the GDP deflators for the domestic country and USA, respectively. (Note that 'standing still' component is another term for change in valuation due to differences in productivity growth).
The decomposition allows one to ascertain the generally ignored component of real exchange rate change--that due to productivity growth differences. The importance of the standing still component can be illustrated with reference to China. For the period 1995-2004, China's excess inflation (with reference to the USA) was of the order of 1.3% per annum; and its nominal exchange rate appreciated by 0.4% per year. So in terms of the direct effect (equation 6c), labor costs in China (again, relative to the USA) increased at 0.9% per year (8) for 10 years or a cumulative 9.4%. But during these years, China grew at an average of 8.0% per capita per year and USA at 2.1% per capita per year. If per capita growth is approximately representative of productivity growth (incidentally, Balassa-Samuelson considerations imply this assumption), then China's labor costs declined at approximately 5.0% per annum. (9) Even if a conservative difference in productivity growth is taken as 4% per annum (rather than 5.9%), it still is the case that the Chinese yuan should have been 34% higher in order for China to have the same competitive level in 2004 as in 1995!
The period 2005 onwards reveals the same pattern. How much is the Chinese yuan under- or overvalued in 2009 with respect to its value in 2005? Computations suggest excess China over US inflation of 1.6% per annum; nominal appreciation of the yuan at an average 3.8% per annum rate; hence, the direct effect is of the yuan appreciating by 5.4 percent per annum. The indirect effect for the same period is 7.2 percent per annum. This is obtained as follows: excess per capita (productivity) growth of 9 percent per annum; RER real income elasticity of 0.8; so a real depreciation of 0.8*9 or 7.2 percent per year. This suggests that the Chinese yuan actually depreciated in real terms at the rate of 1.8% per annum (7.2 - 1.6 - 3.8) over the preceding five years. Perhaps it is not a coincidence that China's growth rate, and the size of its current account surplus, is as large in 2009 as it was during the time its currency was ostensibly appreciating. So in strong contrast to assertions that the Chinese yuan appreciated by close to 20% between 2005 and 2009 (from 8.3 yuan/$ to 6.8 yuan/$), it is the case that China enjoyed a cumulative real depreciation of close to 10% over this period!
Table 2a shows for selected countries the decomposition of the real exchange rate change for two time periods--1980 to 1994 and 1995 to 2009. All computations are with respect to the US dollar. Some important conclusions: First, note the negative and larger share of the standing still component for countries with pronounced currency intervention and related reserve accumulation--the developing, especially East Asian, economies. Second, the developed countries are characterized by very small real change over the 30-year period 1980-2009--this is given by the addition of the (log) effects for the 1980-1994 and 1995-2009 periods. Both Greece and the USA (10) have similar magnitudes of appreciation, 20% and 25% respectively, while Germany and Japan valuation changes are in a narrow range, minus 9% and plus 4%, respectively. Third, that China accomplished between 1980 and 1994 the fastest and largest real devaluation in the postwar period. Over the subsequent 16 years (1995-2009), the Chinese yuan further depreciated, in real terms, by 43% to yield an undervaluation rate of -55.2% in 2009, or an 'equilibrium' close to 4 yuan/dollar. (11)
Table 2b documents regional averages for the two different estimates of changes in currency valuation--the traditional 'direct' estimate and the total change in valuation estimate provided by equations 2 and 3. The direct effect ignores productivity changes; and the difference between the direct and total change reflects the attempts by governments to keep their currencies undervalued by intervening in the foreign exchange markets and accumulating reserves. Estimates of changes are provided for two time periods--1980 to 1994 and 1995 to 2009. The results reinforce the importance of accounting for productivity changes; in the latest 1995-2009 period, the average direct effect for East Asia was small, -5 (log) %; but the total effect was a large -22%. For South Asia the divergence in the two estimates is even greater: the direct effect is 9% overvaluation (mostly due to inflation being higher) while the total effect is 31% undervaluation. The developed economies, as expected, show very little real exchange rate change, and the difference between their direct and total effect is small.
CHINA AND JAPAN--DEJA VU?
The methodological discussion provided in the previous section was necessary to establish the credibility of estimates of real exchange rate change. These estimates can be used to shed light on some important debates relating to current account imbalances and policy invoked changes in nominal exchanges. The question of what exchange rate is appropriate for surplus countries like Japan in the 1980s and China in the 2000s is taken up in this section; the next section examines the relationship between the changing value of the euro and the changing fortunes of the 'weak' European economies.
Despite the record-breaking real devaluations of the Chinese yuan over the last three decades, the desired future course of action is still a two-sided debate. On the one hand there is the historical relationship noted earlier between currency undervaluation and current account surpluses--given that China has large surpluses, one recommended course of action would be to appreciate the currency. On average, each 10% appreciation should help bring down the share of surplus in GDP by 0.3 percentage points. But the Chinese government objects to talk of significant currency appreciation on the grounds that such advice is faulty.
As we all know, Japan faced almost identical inquiries in the mid-1980s and a coordinated government intervention followed to 'forcibly' appreciate the currencies of Germany and Japan. The Plaza agreement catapulted the yen from an average level of around 240 yen/dollar in 1984 and 1985, to 168 yen/dollar in 1986, an appreciation of 40% in just one year! Soon after that revaluation, Japan entered into a long-term decline, the 'lost' decades, and a stagnation it still has to emerge from. Several explanations abound--overly tight monetary policy, the after effects of the breaking of the price bubble in property and stocks, and the emergence of a liquidity trap (Krugman, 1998).
Many believe that the more than doubling value of the yen contributed to growth stagnation. The openly expressed fear of the critics of yuan appreciation today is that China will experience a similar slowdown, and one which will have disastrous consequences for itself and the world economy. As McKinnon and Schnabl (2006) state: 'In the new millennium, history is repeating itself, but now China bashing is superseding Japan bashing' (p. 277). (12) The contention is that if China is made to revalue its currency in any significant manner, history will repeat itself. Will it? This issue--will or can China be the new Japan--deserves investigation.
But first, examination of a related issue: is the Chinese exchange rate undervalued and to what extent? If the undervaluation is large, then is China a currency manipulator? This is more than a decade old question, and an answer is not easy. The US Treasury did reach an affirmative conclusion with regards to China as a currency manipulator in 1994. However, in its 2005 report, the US Treasury distanced itself from that conclusion and described the problems associated with the 'currency manipulator' classification:
'whether countries manipulate the rate of exchange between their currency and the United States dollar for purposes of preventing effective balance of payments adjustments or gaining unfair competitive advantage in international trade is inherently complex, there are several indicators that can provide helpful information in considering this question. However, no single indicator, or set of indicators, in and of itself, can establish that a specific economy has met the technical requirements for designation under the Act, and the context in which these questions are assessed varies with individual country circumstances'.
(US Treasury, 2005, p. 1, emphasis added)
Associated with the concern about manipulation is the (hidden) accusation of mercantilism. Apart from being emotive, mercantilism can mean different things to different people. But in politically incorrect language, the definition of the US Treasury of currency manipulation is mercantilism. The dictionary definition of mercantilism, as defined by Webster, is as follows:
'an economic system developing during the decay of feudalism to unify and increase the power and especially the monetary wealth of a nation by strict governmental regulation of the entire national economy usually through policies designed to secure an accumulation of bullion, a favorable balance of trade, the development of agriculture and manufactures, and the establishment of foreign trading monopolies'.
Substitute unfair competitive practices for 'policies designed to secure an accumulation of bullion', and Webster and the US Treasury might be in agreement.
While the question of currency manipulation is an interesting question in and of itself, the purpose here is to determine whether Japan circa the 1980s is comparable to China circa the 2000s. The comparison is also an obvious one--both large economies, both pursuing the export-led growth model, both alleged to have deeply undervalued exchange rates, and both accumulating reserves. And both causing the USA to have correspondingly large current account deficits. In 1984, the US current account had moved to a historically high deficit of 2.4% of GDP from near balance for most of the preceding decade. And in 2005, USA started moving towards large current account deficits from near parity around the turn of the century.
The Japan then and China today comparison has sense of deja vu. We have seen this movie before and we know what happened. Substitute Japan in the 1980s with China today, and there seemingly is less mystery about the future of China if its currency is allowed to appreciate towards the 'equilibrium'. That, at least, is the view of those who maintain that China's exchange rate should not be 'touched'.
Japan versus China: Testing of the deja vu
The interest in the Japan-China comparison stems largely from a concern about the role of exchange rates on economic growth and on current account surpluses. There are some knowns about the direction of currency changes. Most would agree that a current account surplus 'should' lead to currency appreciation. Most would agree that an undervalued currency should lead to currency appreciation. Unlike the Williamson FEER model which makes surpluses and appreciation near synonymous, the measure of currency valuation proposed above (equations 2 and 3) can and does yield different combinations for the same country. For example, the Indian rupee is deemed undervalued yet it has a current account deficit; the yen is deemed overvalued, yet it has a current account surplus.
This provides the background for the Japan then versus China now comparison presented in Table 3. The experience of the two economies is examined for their respective 5-year time periods of currency misalignment allegations, Japan 1981-1985 and China 2002-2006; as well as their comparative experience for the longer 15-year period, Japan 1970-1985 and China 1990-2005. Unless otherwise stated, the figures in the text represent the shorter 5-year period.
The starting positions for the two countries are very, very different. In 1981, the yen was overvalued by 14%; in 2002, the yuan was undervalued by 36%. The yen stayed overvalued except for the years 1982-1985, and during these exception years it was undervalued by no more than 6%. In striking contrast, the Chinese yuan has been continuously undervalued since 1998, was undervalued by 37% in 2002, by 60% in 2006, and has been outside the extreme 25% band for the last 8 years, not co-incidentally the same time as when the US CAB started to sharply deteriorate.
The cumulative effect of currency undervaluation on income levels is documented in Figure 2 for the 15-year periods 1970-1985 (Japan) and 1991-2006 (China) respectively. The amount of appreciation of the yen is noteworthy, as is the amount of depreciation of the yuan. The differences in income growth are also worth noting. At the end of the 15-year period, per capita income in Japan was log 53.4 or 71% higher; in the corresponding period, China's per capita income was log 135.9 or 290% higher. The currency change may tell the story; the yen appreciated by 35%, the yuan depreciated by a cumulative log 80% or its value in 2006 was less than half (45%) of its real value just 15 years earlier.
[FIGURE 2 OMITTED]
Current account surplus
Japan's current account surplus averaged just 1.9% of GDP in contrast to 5.0% for China. In subsequent years (post-2005) China's current account surplus significantly increased; the surplus as a fraction of GDP averaged over 10% during 2006-2008. This surplus number is large for any non-oil economy, and certainly so for the second largest in PPP terms. When Singapore, or Taiwan, or Korea, or even Japan practiced currency undervaluation, it was something that the world could absorb without significant disruption or imbalances. But when a country with more than a fifth of the world's population practices a beggar thy neighbor policy, it can cause neighbors to have substantially lower growth than would have been otherwise.
Reserve accumulation of both Japan and China seems excessive, but China more so. China's reserves averaged $640 billion a year between 2002 and 2006, (13) a level some 31 times the mean while Japan reserves, at $20 billion, were only about 4.5 times the average for all countries during 1981-1985. In terms of import cover, during its peak year, Japan's reserves accounted for 2 months of imports; in China, reserves were adequate for more than 15 months of imports.
Although the two indicators, undervaluation and current account surpluses, are often associated with mercantilism, there are enough exceptions to suggest that evaluation of mercantilism and currency manipulation is not straightforward. In Bhalla (forthcoming) I offer one method of combining the information independently contained in undervaluation and the share in GDP of the CAB. This mercantilism index is defined as the rank of ranks (named after its mathematician founder Borda as the Borda rank) of the two indicators, that is the rank of the sum of two ranks. Between 1981 and 1985, both Germany and Japan were among the top 11 mercantilist countries. Despite having 'peak' current account surpluses today, both countries have a low mercantilism rank for the period 2002-2006. China ranks 61st in 1981-85, and 8th in 2002-2006. The later years move China even higher in terms of mercantilism. (14)
The mercantilism comparison suggests that the fear that China is in a position to repeat Japan's status is misplaced and inappropriate. In the main, there are two reasons for this optimism. First, that the initial conditions are immensely different. Japanese GDP growth at the time of appreciation of its currency was sharply lower than its own historical average and the average level of Chinese growth in the 2000s. Second, real deflation and slower growth in Japan started in the late 1980s when the yen became overvalued by more than 35% and therefore far beyond the range of 'tolerance'. (15) For China to feel the same overvaluation pressure on growth, the yuan would have to appreciate to a level of 3.25 over the next decade, an unlikely event. Even if the real appreciation is a gradual 5% over the next 10 years, (16) as now appears possible, the Chinese yuan would still be around 25% to 30% undervalued a decade later in 2020. The point simply is that China's rate of currency appreciation has to be significantly faster than the Balassa-Samuelson level of around 4% to 6% per annum to make any dent in the deeply undervalued nature of the Chinese currency.
GDP and export growth
Per capita income grew at only a 2.4% per annum pace for Japan and a level almost four times higher for China, 9.2%. Japanese exports grew at 5.9% per annum while the world (17) median country growth rate was -0.3% during the same period, 1981-1985; by comparison, Chinese exports grew at over 25%, some 11 percentage points above the corresponding median country, during 2002-2005.
Thus, all the data examined--export growth, GDP growth, current account surpluses, mercantilism indices, historical tendencies--indicate only one possibility: the Chinese yuan is much more undervalued today than the yen was at the time of the Plaza agreement. This statement says nothing about the absolute degree of undervaluation; it may even be the case, as some consistently argue (eg Dunaway et al., 2006; McKinnon and Schnabl, 2006; Cheung et al., 2007), that the yuan is absolutely fairly valued or even overvalued. The statement only makes a comment about relative undervaluation; that is if one believes that the yen was undervalued in 1984, then one must believe that the yuan is much more undervalued in 2006.
EURO--HOW DOES IT IMPACT CABs IN THE EUROZONE?
Across the globe from the Far East, there may be a parallel 'mirror image' different peas in the same pod problem--overvalued exchange rates, low growth, and large current account deficits. With the introduction of the euro in January 1999, all 12 countries in the eurozone had the same currency--the euro. These countries, in 1998, ranged in per capita income from US $12,725 in Greece to US $ 26,668 in Germany. This wide range in per capita incomes meant that a competitive euro level for Germany may not have been the same as that for Greece. If not, then individual countries within the euro zone would have differential levels of currency valuation since the RER would differ depending on individual country inflation. Parallel to the China experience of undervaluation helping growth and generating current account surpluses, ceteris paribus, it may be the case that overvaluation leads to a worsening CAB and lower growth.
The euro area can be divided into two parts--Tier I countries comprising of the richer nations (Germany, France, etc) and Tier II countries comprising of the five not-so-rich countries of Greece, Ireland, Italy, Portugal and Spain. The path from currency valuation to current account performance is straightforward. Ceteris paribus, it is expected that currency undervaluation leads to surpluses, and overvaluation leads to deficits. The euro forces the countries to have the same nominal exchange rate, but given different inflation rates, and per capita income levels, these countries have different levels of currency valuation. It is of interest to see if these different valuation levels can help explain the difference in the current account between the Tier I and Tier II countries. Further, given that the euro came into existence in January 1999, a legitimate question arises--how come the Tier II problems were only noticed in 2009/10?
The data are presented for the seven Tier I countries and five Tier II countries as an aggregate in Table 4; the two Tiers are also referred to as the rich and poor countries respectively. Data are presented for selected economic variables for three time periods: Period 1, the 8 years before the euro creation, 1991-1997; Period 2, the 6 years of relatively benign values of the euro 1998-2004; and Period 3, the 4 years of the high euro values, 2005 to 2008. Data for the Tier I and Tier II countries are aggregated to give results for two 'synthetic' economies.
The data strikingly confirm the hypothesized relationship between the value of the euro, and the differential impact on the current account for the poor and rich eurozone countries. At the time of the establishment of the euro in January 1999, the average income of the poor countries was two-thirds of the rich economies. For the prior 7-year period, currency valuation was a near neutral 15% overvalued for the Tier I countries and 32% overvalued for the poorer euro countries where all estimates of valuation are based on equations 2 and 3. Fortunately for the poor economies, the euro continuously declined from its entry level of 1.18 in January 1999 to a low of 0.83 in October 2000. Its average level was 1.13 in 2003. This decline from the entry level of 1.18 helped all the eurozone countries to prosper, and the appreciation of the euro post 2003 to its peak of 1.58 in July 2008, may have been in large part responsible for the problems of the poor economies post 2007. As documented above, the changes in exchange rates and currency valuation directly impact the fortunes in the current account; the economics behind the currency union, and its implications for the rich and poor countries is entertainingly described by Wolf (2010) in his rendition of the grasshopper and ant fable.
For the Tier I countries, no change is observed in GDP growth (constant at 2.1% per annum) and a considerable improvement is obtained in the CAB between the period 1998-2004 (average of 1.7%) and the period 2005-2008 (average of 3.3 %). In contrast, the Tier II countries have much lower GDP growth (1.5% in 2005-2008 versus 2.3% during 1998-2004) and a considerable worsening in the current account (average deficit of 5.9% 2005-2008 versus a deficit of 2% 1998-2004).
Equation 5 postulated a relationship between currency valuation and CAB. The equation or relationship holds for a single country and is easily generalized for groups of countries. Briefly, the hypothesized relationship is as follows with X representing the weighted CAB for time t and region I or II (Tier I or II) and UV the weighted currency undervaluation:
[X.sub.1t] = [a.sub.1t] + b[UV.sub.1t] (7a)
[X.sub.2t] = [a.sub.2t] + b[UV.sub.2t] (7b)
Taking the difference in the two equations, one obtains:
[X.sub.1t] - [X.sub.2t] = ([a.sub.1t] - [a.sub.2t]) + b([UV.sub.1t] - [UV.sub.2t]) (7c)
or d[X.sub.t] = [X.sub.1t] - [X.sub.2t] = ([a.sub.1t] - [a.sub.2t]) + b([UV.sub.1t] - [UV.sub.2t]) = [a'.sub.t] + bdUVt (7d)
Equation 7d represents a fixed 'region' effects model. It is estimated by regressing not the levels but rather the first differences of the relevant variables. The difference model allows for the control of effects which are particular to a country or region, for example Germany as a perennial surplus country possibly due to its high taste for savings, ceteris paribus.
A simple regression for annual data for the period 2000-2008 between dX and dUV yields a correlation of 0.95 and a coefficient on dUV of minus 0.74; for the longer period 1980-2008, the correlation is still a large 0.48 and the coefficient is -0.50. These are large effects; for the most recent post-euro, post-1999 period, these estimates indicate that if the Tier II economies were to depreciate their currencies relative to the Tier I economies (and thereby the rest of the world) by 10%, their current account imbalance would improve by 7.4 percentage points, that is move into a surplus of 1.8% of GDP from the average deficit level over these 9 years of minus 5.7% of GDP. In the pre-euro period 1991-1998, the current account of the Tier II economies was approximately in balance (Table 4).
Figure 3 plots the two difference variables against each other, that is the difference in the current account balance between the Tier I and Tier II eurozone countries against the difference in their levels of valuation. It is a near perfect fit; the strength of the correlation underlines the importance of currency valuation as a determinant of current account imbalances and provides a plausible explanation for the troubles in the Tier II countries post the appreciation of the euro since 2004.
[FIGURE 3 OMITTED]
This paper has attempted to document the symmetric effects of currency valuation on economic growth and imbalances in the current account. It is hypothesized that if a currency is undervalued, it helps economic growth to proceed faster than otherwise, and for current account surpluses to be larger than otherwise. Contrasting experiences of as diverse economies as China, Japan, and the eurozone post the introduction of the euro in 1998 are evaluated from the prism of currency valuation.
This paper suggests a new measure of currency misalignment--this. measure is based on a hypothesized S-shaped relationship between the RER and (log) real income levels. This specification allows for better measurement of currency valuation--and the relationship between currency valuation and growth to be more firmly grounded. Each 10% increase in undervaluation leads to an increase in per capita growth rates of around 0.1 percentage point. If a country has large undervaluation levels, for example China, this means a sustained extra per capita growth of around 0.5% per year.
The large currency undervaluation in China is defended by many on the grounds that revaluing the currency would lead to a Japan-type crisis. A detailed comparison of Japan in the 1980s and China in 2000s suggests that this fear is misplaced. The primary reason being that the Japanese currency was not undervalued by much at the time of the Plaza agreement, and that the Chinese yuan is deeply undervalued today. A 5%-10% real appreciation of the yuan per year is both needed and unlikely to be disruptive to the growth process. Such a revaluation will reduce the growth rate but not by much. Given the higher productivity growth in China, a 3-4% appreciation per year is needed just to keep the valuation levels constant.
The role of currency valuation in affecting current account imbalances is reinforced by the analysis of deficits in the eurozone, in particular among the overvalued Tier II economies of the European Union, for example Greece, Portugal, Spain, etc A very strong relationship is observed between the changes in overvaluation of these Tier II economies and the changes in their current account deficits in the last 10 years.
The findings of this paper reiterate the fact that currency misalignments have consequences, and that large misalignments have large consequences. It is quite obvious that correction of currency misalignments is a necessary first step towards a less disharmonious economic and political future.
Today and tomorrow dictate that the major realignment of the US dollar has to come with respect to ex-Japan Asia, especially China. China has a misalignment in currency as large as 60%, and a misalignment increasing by 3-5% per year due to the higher productivity growth or 'standing still' argument. Never has a currency been so misaligned for so long; never, obviously, with such a large mass of population. Add to it the need by China's Asian neighbors, including Japan, to stay competitive. The problem becomes worse. The problem is no more the American consumer and the fact that US savings are low.
A common perception is that one needs a policy agreement, a la the Plaza agreement, to achieve a large change in the level of a currency. Not so. Indeed, more often than not, developing countries have achieved a substantial real devaluation, a substantial reduction in costs, by simply keeping the exchange rate fixed. This is counter to the Balassa-Samuelson effect, according to which a currency should appreciate with fast growth. In recent years, this effect has been thwarted by countries intervening to keep their exchange rates 'competitive'. The lead was provided by China, and followed by other, especially Asian, countries. The effect of intervening, keeping the currency cheap, and not having much excess inflation has meant that the Asian currencies have become even more undervalued. This 'standing still' real depreciation is a large part of the currency story post the East Asian crisis.
Speculation about the future
At present, there is the perception of an important change within China. It no longer is a surprise, or is shocking, for Chinese officials and Chinese think tank academics to argue that the present and future is different than the past--at least as far as Chinese exchange rate policy is concerned. In the past, the arguments were as follows. In 1997, at the onset of the East Asian currency crisis (a crisis whose origins most likely were in the large Chinese devaluations over the period 1989-1994), China claimed that as a gesture of solidarity with the US and the global economy, it would not devalue in retaliation to the East Asian devaluations. Later, China argued that since it was a poor country with large underemployment, (18) it needed an undervalued currency to grow fast. Next it argued that despite huge trade surpluses, its financial system was very weak and could not take a currency adjustment. It then admitted that it did need to increase consumption among the poor but no, it could not appreciate the currency because that would hurt the farming sector--but the farming sector incomes would benefit from cheap imports.
The chances that China begins to revalue its currency is high. The global players, for example G20, have begun to identify problem spots in this post-crisis world so that a 2008-type deep crisis does not visit us again. A post-crisis currency realignment is something all players recognize as inevitable. As inevitable as the Balassa-Samuelson effect. Before 2008, the world was also slightly different. This alignment of natural interests did not occur before because we did not realize how interdependent we were. It wasn't allowed to happen before, but the future is likely to be different. Going forward, it is likely that the yuan will appreciate, in real terms, by about 5-10% a year for the rest of this decade.
Forecasting the future of the euro vis-a-vis the Tier II economies is a lot more difficult. The value of the euro involves the actions of several nations and not just one nation. Compounding the problem is that the countries comprising the euro are part, and yet not part, of a political union. One solution would be for the Tier I economies of Europe to significantly increase their wages, at least relative to the wages of the Tier II economies. Not surprisingly, an increase in wages, and therefore domestic consumption, and an increase in imports relative to exports, is the same peas in the pod solution that will work towards appreciating the yuan and reducing its current account surplus.
I would like to thank Paul Wachtel for helpful comments and suggestions.
Baily, MN and Lawrence, RZ. 2006: Competitiveness and the assessment of trade performance. In: Mussa, M (ed). C. Fred Bergsten and the World Economy. Peterson Institute for International Economics: Washington.
Balassa, B. 1964: The purchasing power parity doctrine: A reappraisal. Journal of Political Economy 72: 584-596.
Bergin, P, Glick, R and Taylor, AM. 2004: Productivity, tradability, and the great divergence. March. Available at: http://elsa.berkeley.edu/~obstfeld/281_sp04/amt.pdf.
Bhalla, SS. 2006: Second among equals: The middle class kingdoms of India and China, executive summary. Peterson Institute for International Economics: Washington, June.
Bhalla, SS. 2007: Second Among Equals: The Middle Class Kingdoms of India and China. May. Available at: www.oxusinvestments.com.
Bhalla, SS. 2008: Economic development and the role of currency undervaluation. The Cato Journal 28(2): 313-340.
Bhalla, SS. forthcoming: Devaluing to prosperity: Misaligned currencies and their growth consequences. Peterson Institute of International Economics: Washington.
Cheung, Y-W, Chinn, MD and Fujii, E. 2007: The overvaluation of renminbi undervaluation. Journal of International Money and Finance 26(5): 762-785.
Clark, G. 2007: Farewell to alms. A brief economic history of the world. Princeton University Press: Princeton.
Dollar, D. 1997: Outward-oriented developing economies really do grow more rapidly: Evidence from 95 LDCs, 1976-1985. Economic Development and Cultural Change 40(3): 523-544.
Dunaway, S, Leigh, L and Li, X. 2006: How robust are estimates of equilibrium exchange rates: The case of China. IMF Working Paper 06/220, Washington, DC.
Easterly, W. 2005: National policies and economic growth: A reappraisal. In: Aghion, P and Durlauf, S (eds). Handbook of Economic Growth, Chapter 15. Elsevier: Amsterdam.
IME 2010: World Economic Outlook database, April. International Monetary Fund: Washington.
Johnson, S, Ostry, JD and Subramanian, A. 2007: The prospects for sustained growth in Africa: Benchmarking the constraints. IMF Working Papers 07/52, International Monetary Fund: Washington, March.
Krugman, P. 1998: It's Baaack: Japan's slump and the return of the liquidity trap. Brookings Papers on Economic Activity 2: 137-187.
McKinnon, R and Schnabl, G. 2006: Devaluing the dollar: A critical analysis of William Cline's case for a new Plaza agreement. Journal of Policy Modeling 28 (6): 683-694.
Rodrik, D. 2008: The real exchange rate and economic growth. Brookings Papers on Economic Activity 2: 365-412.
Rostow, W. 1960: The stages of economic growth: A non-communist manifesto. Cambridge University Press: Cambridge.
Samuelson, PA. 1964: Theoretical notes on trade problems. Review of Economics and Statistics 46: 145-154.
US Treasury. 2005: Appendix: Analysis of exchange rate policies pursuant to the act in report to Congress on international economic and exchange rate policies. Washington, DC.
Williamson, J (ed). 1994: Estimating equilibrium exchange rates. Institute of International Economics: Washington.
Wolf, M. 2010: The grasshoppers and the ants - A contemporary fable. Financial Times, 26 May.
Woodford, M. 2008: 'Is an undervalued currency the key to economic growth', comment on Dani Rodrik, 'The Real Exchange Rate and Economic Growth'. Brooking Papers on Economic Activity 2: 420-437.
SURJIT S BHALLA
Oxus Investments, S-160 Panchshila Park, New Delhi 110017, India.
(1) The US current account was in surplus during the early 1960s, averaging around 0.4% of GDP.
(2) Penn World Tables PWT 6.1 is the primary source for data on PPP incomes; the data have been extrapolated to 2009 using inflation, exchange rate, and real income growth data from IMF WEO (2010).
(3) While several explanations have been offered to explain the acceleration of India's growth rate to about 5.5% in the 1980s, Bhalla (forthcoming) finds that the reallocation process can explain virtually all of the acceleration from around 3.5% in the early 1960s to 5.5% in the 1980s.
(4) Nominal PPP 1996 income for years after 2000 are obtained from IMF and World Bank (World Development Indicators) and linked to the Penn data. Conversion into real values is done via the US GDP deflator. The starting year was chosen as 1996 because the base year for estimates of PPP (Penn Tables) is 1996.
(5) In Bhalla (forthcoming) results for different models are presented; regardless of the specification, or period selection, or method of estimation a very strong relationship between currency valuation and growth is obtained. Further, the results suggest a symmetric effect--currency overvaluation hurts growth, undervaluation helps growth, by near equal amounts.
(6) See Baily and Lawrence (2006) for a detailed analysis of this relationship for the USA; their specification has the log of the ratio of exports to imports as the dependent variable.
(7) The BIS has recently released REER data for over 60 countries for the time period 1994-2010.
(8) All percent changes are in log terms.
(9) Obtained as the difference in the per capita growth rates and the direct effect REER, that is 8.0 minus 2.1 gives a relative productivity growth of 5.9%. Labor costs rose at 0.9% per year, so net labor cost advantage to China was of the order of 5% per year.
(10) All valuation calculations are done with respect to the US dollar. Valuation for the USA is obtained as the negative of a weighted valuation of 37 countries which comprise the Fed's BROAD index of the US dollar.
(11) The calculation is as follows. In 2009, the average level of the yuan/US dollar rate was 6.83. A log undervaluation level of -55.2% yields an equilibrium level of 3.94.
(12) It gets a bit confusing because many of the critics who fear a Japan repeat also believe that the RER, and therefore the under- or overvaluation of a currency, does not affect economic growth.
(13) In March 2007, China's reserves exceeded $1.2 trillion and were thus almost twice the average over the preceding 5 years.
(14) Hong Kong manages to have a single digit rank in both periods. South Korea does not appear to have played unfairly in either period (rank of 45 and 17 for the earlier and later period, respectively). In contrast, Taiwan is one of the most mercantilist countries during the naughties (rank 6).
(15) The yen reached a peak overvaluation level of log 66.6 %, or actual 93%, in 1995 when the yen averaged 94.1 for the year.
(16) It is worth emphasizing that the Goldman Sachs BRICS report assumes, in its rather bullish forecasts, a real appreciation of the Chinese yuan at the rate of 3.5% per year. One of the flaws in the report is that it did not recognize the dependence of Chinese high growth on a depreciating real currency.
(17) The world calculation refers to about 90 countries with population greater than 1 million and excluding the major oil-exporting economies.
(18) By implication, the rest of the developing and developed world have to have a lesser unemployment problem, ceteris paribus.
Table 1: Investment and growth (1950-2009): The importance of currency undervaluation Log (initial) income Undervaluation in per capita initial year Investment (%GDP) Fixed effects -2.00 *** -0.046 *** (-2.6) (-7.65) GMM -5.05 -0.038 *** (-1.3) (-3.92) Per capita growth (%) Fixed effects -3.65 *** -0.007 *** (-10.8) (-2.78) GMM -8.76 *** -0.014 *** (-3.06) (-2.26) CAB (% of GDP) Fixed effects (1950-2009) -0.03 *** (-5.02) Fixed effects (1980-2009) -0.25 *** (-3.43) Adjusted [R.sup.2]/Wald [chi square] Number of obs. Investment (%GDP) Fixed effects 0.4821 978 GMM 191.2 678 Per capita growth (%) Fixed effects 0.2734 1186 GMM 571.47 721 CAB (% of GDP) Fixed effects (1950-2009) 0.4078 871 Fixed effects (1980-2009) 0.4435 687 Notes: (1) CAB--Current Account Balance as % of GDP. (2) *** p <0.01, ** p <0.05, * p <0.1, t-statistic in parentheses. (3) Regressions are based on a pooled cross-section of 5-year periods; only countries with at least 30 years of data (115 countries) were selected. (4) Undervaluation is measured as a residual from equation 3, that is undervaluation is equal to RER-RER* where RER* is the predicted value from equation 3. Source: Penn Tables 6.1, IMF WED database, April 2010, authors calculations Table 2a: Direct and indirect (standing still) estimates of (log) changes in real exchange rates Country 1980-1994 [RER.sub.direct] [RER.sub.sstill] Total (1) (2) (3) China -127 -58 -185 India -63 -44 -108 Turkey -82 -18 -100 Indonesia -43 -55 -99 Korea 16 -61 -46 Saudi Arabia -55 29 -26 Argentina -27 5 -22 Australia -17 -2 -20 France -10 1 -9 Germany 0 -8 -8 Canada -9 2 -8 Brazil -7 1 -6 Singapore 26 -32 -6 UK 2 -7 -5 Italy 12 -8 3 Greece 4 4 8 Mexico 13 -4 9 USA 0 10 10 South Africa 6 7 13 Russia 33 -4 29 Japan 48 -11 37 Country 1995-2009 1980-2009 [RER.sub.direct] [RER.sub.sstill] Total Total (1) (2) (3) (3) China 45 -88 -43 -228 India 7 -73 -66 -174 Turkey 41 -25 16 -84 Indonesia 21 -39 -18 -117 Korea -33 -15 -48 -94 Saudi Arabia 28 6 33 7 Argentina -58 -10 -68 -90 Australia 20 -3 16 -3 France 9 0 9 -1 Germany -5 3 -1 -10 Canada 16 -3 13 5 Brazil 19 -16 3 -3 Singapore -19 -1 -20 -26 UK 7 -5 2 -3 Italy 25 2 27 30 Greece 28 -16 12 21 Mexico -2 -3 -5 4 USA 0 15 15 25 South Africa -5 -16 -20 -8 Russia 71 -31 40 69 Japan -38 4 -33 4 (1). Currency undervaluation estimates obtained from equations 2 and 3 in text. (2). The direct effect is the conventional estimate of change in the real effective exchange rate (REER), that is due to exchange rate fluctuations and inflation differences. The indirect effect (column 2) refers to the change in the real exchange rate due to the differences in productivity growth--the 'standing still' effect. (3). Countries are ordered according to the size of the real devaluation in 1980-1994. Table 2b: Different estimates of change in currency valuation Region Estimate of RER change Direct only (in log %) 1980-1994 1995-2009 Developed economies -1 12 East Asia -12 -5 Russia and Eastern Europe 27 76 Latin America -8 13 Middle East + North Africa -28 16 South Asia -31 9 Sub-Saharan Africa -57 35 All regions -18 24 Region Estimate of RER change UV Direct + Standing Still (log %) 1980-1994 1995-2009 Developed economies -7 8 East Asia -29 -22 Russia and Eastern Europe 30 39 Latin America -15 1 Middle East + North Africa -23 10 South Asia -60 -31 Sub-Saharan Africa -29 1 All regions -16 3 Notes: The direct effect refers to the 'traditional' currency change and inflation differential computation of real exchange rate change; the 'direct plus standing still effect' refers to the total change in valuation which is the log sum of the direct effect plus the Balassa-Samuelson indirect effect which incorporates differences in productivity growth. Table 3: Japan and China--Not a case of deja vu 5-year period Japan China 1981-1985 2002-2006 Growth (in %) GDP 3 9.7 per capita GDP 2.4 9.2 Exports 5.9 25.3 Imports -0.7 22.9 Reserves 1.6 32 Average levels Reserves ($, bil) 26 641 Currency valuation--DYP (in %) -0.5 -50.2 Currency valuation--JOS (in %) 20.1 -55.9 Mercantilism (Rank in 1985/2006) 15 4 Mercantilism Rank (Average) 9.6 8.5 Levels as % of GDP Current account balance 1.9 5 Savings 31.3 46.7 Household consumption 54.4 38.8 Investment 29.5 41.7 15-year period Japan China 1970-1985 1991-2006 Growth (in %) GDP 4.3 9.7 per capita GDP 3.3 8.9 Exports 9.7 18.2 Imports 8 18.2 Reserves 13.2 22.4 Average levels Reserves ($, bil) 20 272 Currency valuation--DYP (in %) -1.8 -17.9 Currency valuation--JOS (in %) 14.1 -45.7 Mercantilism (Rank in 1985/2006) Mercantilism Rank (Average) 9.2 16.3 Levels as % of GDP Current account balance 1 2.6 Savings 33.8 42.5 Household consumption 52.4 42.5 Investment 33.2 39.7 Notes: (1) For definition of mercantilism index, see text. (2) A negative sign on currency valuation means undervaluation. (3) DYP and JOS refers to Bhalla (forthcoming) or equation 3 in the text and Johnson-Ostry-Subramaniam, respectively. Table 4: The eurozone countries--current account balances and related variables Tier I countries Currency Current account Average growth valuation balance in GDP in % % of GDP in % 1 2 3 Period 1.1991-1998 32 0.3 1.8 2.1999-2004 13 1.7 1.9 3.2005-2008 32 3.5 2 Tier II countries Currency Current account Average growth valuation balance in GDP in % % of GDP in 4 5 6 Period 1.1991-1998 15.3 -0.1 1.8 2.1999-2004 0.5 -1.9 2.6 3.2005-2008 26.8 -5.8 1.7 Notes: (1) TIER II refers to the countries Portugal, Ireland, Italy, Greece and Spain; TIER I are the other eight countries of the original eurozone, for example Germany, France, Netherlands, etc. (2) The currency valuation estimates are obtained for the eurozone countries on the basis of their per capita income and exchange rate prevailing at the time of their entry into the eurozone in January 1998. Individual country inflation rates are used to estimate the nominal and real exchange rates.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Symposium Article|
|Author:||Bhalla, Surjit S.|
|Publication:||Comparative Economic Studies|
|Date:||Sep 1, 2011|
|Previous Article:||Emerging economies perspectives.|
|Next Article:||Crisis and growth in the advanced economies: what we know, what we do not, and what we can learn from the 1930s.|