Truths and myths about RMB misalignment: A meta-analysis

We conduct a meta-regression analysis of 69 studies that generated 937 renminbi (RMB) misalignment estimates. The Bayesian Model Averaging (BMA) approach is adopted to allow for model selection and sampling uncertainties in assessing effects of study characteristics on these RMB misalignment estimates. Misalignment estimates are found to be influenced by the eight selected study characteristic types in our median probability model. The RMB misalignment estimate from models with various hypothetical combinations of study characteristics, however, is mostly insignificantly different from zero. It is also shown that the set of significant study characteristics is sensitive to the use of the least squares estimation method and the choice of benchmark study characteristics.