How Robust Are Estimates of Equilibrium Real Exchange Rates: The Case of China

Abstract

Abstract. Assessments of a country's real exchange rate relative to its 'equilibrium' value as suggested by 'fundamental' determinants have received increasing attention. Using China as an example, the present paper illustrates models commonly used to derive equilibrium real exchange rate estimates. The large variance in the estimates raises serious questions about the robustness of these results. The basic conclusion is that, at least for China, small changes in model specifications, explanatory variable definitions, and time periods used in estimation can lead to very substantial differences in equilibrium real exchange rate estimates. Therefore, such estimates should be treated with great caution.

Renminbi's misalignment: A meta-analysis

“…The amplitude of the deviations in Chinese currency compared to its fundamentals is very large; most studies show a significant misalignment of Chinese currency that ranges from -66% to almost 162%. Dunaway, Leigh and Li (2006) show that this misalignment range is driven by numerous factors, including real equilibrium exchange rate models, econometric methodologies, samples of countries, choices of variables, methods of construction and the data's frequency.…”

A Simple Model and its Application in Currency Valuation

“…The uncertainty of econometric specification for the Penn effect model was also confirmed by Cheung et al (2010). Besides the Penn effect model, the fault also appears in all other models that use an econometric method, such as the BEER and the macroeconomic balance models; see Dunaway et al (2009).…”

A Comparison of the Beer and Penn Effect Models Via Their Applications on the Valuation of the Renminbi

“…Dunaway et al (2006) have proven that, at least for China, small changes in model specifications, explanatory variable definitions, and periods used in estimation can lead to very substantial differences in the equilibrium real exchange rate. This can also be seen from Table 2 in Section 5.3.1, in which the RMB misalignments in each same year derived by different economists are different regardless of whether the economists use same or different kinds of models.…”

“…The Penn effect model is named as "the (bivariate) relationship between relative price and relative output levels" by Cheung et al (2007). The Penn effect model and the BEER model are not differentiated by the studies of Dunaway and Li (2005) and Dunaway et al (2006), where the two models are both named as "the extended PPP approach. "…”