Wavelet Analysis and the Forward Premium Anomaly

摘要: Forward and corresponding spot rates on foreign exchange markets differ so that forward cannot be used as unbiased predictors for future rates. This phenomenon has entered the literature under heading of Premium Anomaly. We argue standard econometric analyses implicitly assume relationship is time scale independent. use wavelet analysis to decompose rate changes, premia, using maximal overlap discrete transform (MODWT). Then we estimate a scale-by-scale basis, thereby allowing market inefficiencies such noise, technical, feedback trading well fundamental rational trading. The results show premia serve changes (unbiasedness hypothesis) certain scales only. Monthly weekly data concerning Euro, US-dollar British Pound periods from 1 month 5 years analysed. find unbiasedness hypothesis rejected if reconstructed medium-term long term components. most prevalent in transaction up year.