Revisiting the PPP puzzle: Nominal exchange rate rigidity and region of inaction

The paper revisits the purchasing power parity (PPP) puzzle and investigates potential sources of rigidity in the real exchange rate convergence. We propose two hypotheses on the slow convergence estimates in previous studies, and re-estimate the convergence rates to test the hypotheses. First, the slow convergence estimation is potentially from the forex market intervention by the central banks which claim to freely float their exchanges. Second, the real exchange rates may not show any convergence if they are already converged and within the “region of inaction.” We identify the regions of inaction (1) by creating bands around averages of real exchange rates, assuming that the PPP forms around the long-run averages, and (2) by using a Bayesian Dynamic Linear Model (DLM), assuming that real exchange rates converge to the different levels over time. We use a standard fixed effect estimation as well as a mean group panel estimation to handle aggregation biases in standard panel estimation. The results confirm the propositions and make the puzzle less puzzling; we find approximately 30%–65% shorter half-life estimates than the standard fixed effects estimates in previous studies.