Nguyen, B. H. (2010). Cross sectional dependence and PPP [Master Thesis, Technische Universität Wien]. reposiTUm. https://resolver.obvsg.at/urn:nbn:at:at-ubtuw:1-50398
Panel unit root methods have been widely used in studying the purchasing power parity (PPP) hypothesis. The recent study of Imbs et al. (2005), using first generation panel unit root tests, finds evidence for PPP. Based on stationarity, they argue that the failure to allow for sectoral heterogeneity induces a positive bias in persistence estimates. However, firstly, their panel unit root tests used do not take into account cross-sectional dependence, a phenomenon pervasive in real exchange rates. Secondly, they do not provide the calculation for the bias when some of the series in the panel are integrated. Consequently, a reassessment will take place based on the approach outlined in Wagner (2008). Our first finding is that once we account for cross-sectional dependence, the evidence favoring PPP disappears. Our second contribution is to show that the pooled OLS estimator is biased upwards in cross sectionally independent mixed panels.
