Multiple Testing of the Forward Rate Unbiasedness Hypothesis Across Currencies

We develop exact distribution-free test procedures for joint inference about the forward rate unbiasedness hypothesis (FRUH) across multiple currencies. The procedures can be applied with either levels or differences specifications. This unified approach proceeds with sign and signed rank tests for each currency and then uses Monte Carlo resampling to control the overall Type I error rate of either: (i) global FRUH tests obtained via combinations of the p-values; or (ii) individual FRUH tests using multiplicity adjusted p-values. Our framework allows for missing data and for the presence of time-varying conditional covariances between currencies. The usefulness of the new procedures is illustrated with a simulation study and with assessments of the FRUH across 13 currencies in an unbalanced panel. Multiplicity adjusted p-values reveal that the joint FRUH rejections are primarily driven by just a few of the more minor currencies