This implies that real exchange rate dynamics are far more complex than assumed by previous
studies which have based their analysis on standard unit root tests and either reject PPP or report very
low rates of mean reversion. As pointed out by Sarno and Valente (2006) and Taylor et al. (2001), these
tests also have low power because they apply common factor restrictions to the dynamics of prices
and nominal exchange rates and neglect nonlinear adjustment dynamics. Providing estimates for the
half-life times of shocks based on a model which simultaneously allows for nonlinear adjustment
without implying such restrictions, this study confirms previous results by Sarno and Valente (2006)
and Taylor et al. (2001) that the half-life times of real exchange rate shocks are significantly below the
consensus benchmark of 3–5 years.