Using Survey Data to Test Market Efficiency in the Foreign Exchange Markets.

This paper uses cointegration methods to test the market efficiency hypothesis (MEH) in the foreign exchange markets. Four exchange rates are considered--all relative to the U.S. dollar: British Pound, Deutsche Mark, Swiss Franc and Japanese Yen. Survey data on expectations are used to see whether the violation of the MEH is due to expectational errors or risk premia. The results differ for the one-week ahead and the one-month ahead forecasts. With the weekly data they conclude that it is risk premia, and with the monthly data it is both expectational errors and risk premia that account for the violation of the MEH. Given the volatility of the exchange markets, it appears that forecasts over an extended period fail tests of rationality, but one-week ahead forecasts do not fail such tests.