Doubts as to the Question of Renminbi Undervaluation
In the period leading up to the July 2005 announced change in exchange rate policy, debate over the extent to which economic conditions justified exchange rate adjustment in China reached crescendo pitch.
The needed degree of revaluation, depending upon the source and the particular estimation method employed, was seen as ranging anywhere from zero to 50% or more. Much analysis focused on purchasing power parity, the premise that, after converting domestic currency into foreign currency at the going exchange rate, internationally traded goods should be purchasable for essentially the same price abroad as at home. Empirical application of this premise is complicated, however, both by the need to capture prices on a broad range of goods and by the need to abstract from the contribution of “nontradables,” that is, such factors as land and labor inputs that, even if they differ substantially in price across countries, cannot be readily or legally shipped from place to place. Although the narrow focus on a single item under The Economist’s popular “Big Mac Index” hardly addresses these requirements, its transparency and popularity have given considerable prominence to its implication that the renminbi was dramatically undervalued in the period preceding the July 2005 policy adjustment.The Big Mac Index compares the price at McDonald’s restaurants abroad to the US price, and assesses whether dollars converted into local currency at the current exchange rate would have more or less purchasing power than in the United States. The Big Mac has tended to be considerably cheaper in dollar terms in mainland China, implying that, by 2003, revaluation of over 50% would have been needed to equalize prices across the two countries. As pointed out by Yang (2004), however, most of the Big Mac’s price is accounted for by nontradable local services like labor, rent, and electricity, which are substantially cheaper in China than in the United States.
Allowing for these cheaper nontradable components could quite easily account for the apparent price discrepancy, even assuming full equality in the portion of the price reflecting tradable components like the meat, the bun, and the paper. Meanwhile, The Economist’s own alternative “Tall Latte Index” showed the dollar-equivalent Tall Latte price at Starbucks to be, over essentially the same time period, nearly identical across the two countries - costing $2.80 in the United States and requiring $2.77 to make the same purchase in China (Max, 2004). The greater equality in price for the Tall Latte may well reflect a more important role for tradable components like (relatively more expensive) coffee beans but, more importantly, it suggests that considerable caution needs to be applied in inferring renminbi under- or overvaluation from product-specific price comparisons.The “Apple Index” compiled by Laurenceson and Tang (2006), based on the price of Apple Inc.’s globally available electronic products like iMac and iPod, not only focuses on products with a high value share for tradable components (principally flash memory) but also attempts to control for tariffs and taxes and for the value of the residual nontradable inputs. This “Apple Index” consistently suggests closer approximations to purchasing power parity than the Big Mac Index across a broad range of countries both in Asia and elsewhere. And there is no support for renminbi undervaluation. Moreover, Laurenceson and Tang (2006) show that the Apple Index does not share the Big Mac Index’s tendency toward higher degrees of exchange rate undervaluation at lower national per capita income levels. To the extent that nontradable goods are cheaper in lower income countries, the prices of goods with substantial nontradable value components will be biased downwards in such cases and produce erroneous estimates of exchange rate undervaluation - estimates that should be based only on tradable goods.[13] Some, if not all, of the extreme renminbi undervaluation suggested by the Big Mac Index simply reflects the operation of this bias.
Among broader price indices, the World Bank’s International Comparison Program is generally considered the best source of data for purchasing power parity comparisons. However, even this series features a strong inverse correlation between the degree of implied exchange rate undervaluation and per capita income levels similar to that seen forthe Big Mac Index (Funke and Rahn, 2005, pp. 467-468). A considerable number of studies have recently sought to extend the purchasing power parity approach by incorporating the effects of nontradables via a relative productivity measure. This is based on the premise that cheaper nontradable goods in countries like China reflect the relatively lower wages that would tend to follow from lower productivity levels. As relative productivity rises, therefore, wages and nontradable goods prices should tend to catch up with more advanced economies, implying a higher real exchange rate. Recent time series analysis, incorporating alternative sets of additional macroeconomic control variables, suggests estimated degrees of undervaluation ranging from minimal (Shi and Yu, 2005; Goh and Kim, 2006) to a moderate 11-12% (Funke and Rahn, 2005) through the end of 2002.[14] Shi and Yu (2005, p. 43) do point to a recent rising trend, however, with the estimated degree of undervaluation increasing from an average of just 4.24% from the third quarter of 1999 through the first quarter of 2002 to 10% over the period from the second quarter of 2002 to the third quarter of 2004.
Still another approach is to incorporate the effects of relative productivity levels by employing cross-sectional analysis to assess the “average” relationship between relative productivity levels, or relative per capita income levels, and real exchange rates across countries. Frankel (2006) suggests that, based on the extent to which China’s real exchange rate failed to keep up with the “average” reaction to income growth, the renminbi was undervalued by around 35% in 2000.
Using a similar technique, Coudert and Couharde (2007, p. 574) suggest undervaluation in at least the 44-46% range for 20002004.[15] Controlling for other factors, including sampling uncertainty, and considering trends over time as well as across countries, Cheung, Chinn, and Fujii (2007b) conclude, however, that such strong findings of renminbi undervaluation are not robust. Meanwhile, Dunaway, Leigh, and Li (2006) show that relatively small changes to the sample period, specific set of variables included, or particular set of countries included, tend to have disproportionately large effects on the findings, producing variations of up to 50 percentage points in the implied deviations of the exchange rate from its equilibrium value. Extended purchasing power analysis, of the type described earlier, is also seen to be highly sensitive to relatively small changes in the setup of the empirical test.Dunaway, Leigh, and Li’s (2006, p. 9) conclusion that “estimates of a country’s equilibrium real exchange rate need to be treated with a great deal of caution” is certainly worth bearing in mind when confronted with overly-strong statements implying that a specific degree of renminbi undervaluation is a proven fact (cf. Bergsten, 2007). Although it is true that the available empirical studies do consistently point to a movement in the direction of renminbi undervaluation in recent years, this must be placed in the context of likely significant overvaluation in the aftermath of the 1997-1998 Asian financial crisis. Examination of the historical record seems to offer further reason for being wary of the case for very large renminbi revaluation. As Figure 1.2 shows, China’s effective real exchange rate in 2006, far from being unprecedented, represented nothing more than a return to the levels seen during 1996. Indeed, the effective real exchange rate actually trended slightly upward during 2005-2006. This occurred in the midst of gradual appreciation of the nominal renminbi exchange rate against the US dollar, strengthening from 8.28 RMB/$US on July 20, 2005 to 7.62 RMB/$US by June 30, 2007. Rising reserve inflows and Chinese trade surpluses may well justify continued appreciation. However, neither historical comparisons nor recent empirical evidence seems to offer any real, consistent support that more drastic moves are justified by Chinese economic fundamentals - rather, recent political pressure seems to be driven more by concerns about US and global imbalances, as discussed in Chapter 2.