Exchange Rate Dynamics and the Relationship between the Random Walk Hypothesis and Official Interventions

This paper examines the empirical evidence that official interventions are associated with periods of high predictability in exchange rate markets. We employ a block bootstrap methodology to build critical values for the Variance Ratio statistics and test for predictability within moving windows of fixed length sizes for major developed countries currencies. Empirical results suggest that interventions are indeed associated to periods of increase in predictability and that time varying risk premium may, at least partially, explain such results.