On testing the hypothesis that premium in US$/$A forward exchange rate is constant: a signal extraction approach

In this paper, we implement the signal extraction approach, a methodology from engineering literature, to identify and measure premia in the pricing of 30-day and 90-day US$/$A forward foreign exchange. The estimated premium models indicate that the variance of the premium term is not equal to zero and reject the hypothesis that the premium is constant. It also lends support to some other findings that the variance of the premium term accounts for more that 70 per cent of the variation in forecast errors that results from the use of current forward rates as a predictor of future spot rates.