Empirical studies have found that many foreign exchange markets are `efficient', but this result is usually sensitive to the specific method of testing market efficiency, which tends to be less apparent for tests based on series of differences between forward and spot rates than for those based on their levels.

In Discussion Paper No. 627, Juan Ayuso, Research Fellow Juan J Dolado and Simón Sosvilla-Rivero investigate the validity of the efficient markets hypothesis for Spain during 1985-91. They use daily data to test the efficiency of the peseta forward market against the dollar, using both levels and differences in the forward series relative to the spot rate. This in effect tests for agents' preparedness to pay a risk premium by resorting to the forward market rather than risking capital losses that arise from currency appreciation or depreciation if they leave their foreign currency positions open.

Their results indicate that there are such risk premiums in the one- and three-month segments of the peseta/US dollar forward market. On average these are small, however, amounting to 1.2% and 3.75% of the expected future spot rates respectively; this 1:3 ratio suggests a possible linear relationship between the premium and the time needed to resolve the uncertainty associated with holding the position open. Their estimates based on the premiums' volatility are much less conclusive, but they appear to have smaller values than the (rational) prediction errors, which are the other main source of volatility in the model. The authors therefore reject the hypothesis of efficiency in the strictest sense, but their results indicate only negligible departures from efficiency.

Ayuso, Dolado and Sosvilla-Rivero also suggest that the discrepancy often found between the results of tests of forward market efficiency based on series of their levels and those based on deviations from current spot rates may arise because the proximity of identically dated forward and spot rates introduces bias and tends to inflate the standard deviations of the estimated coefficients for data based on deviations. Replacing the current spot rate with the spot rate ruling in a previous period can resolve this problem and eliminate the contradiction between the results based on different types of data.

Efficiency in the Peseta Forward Exchange Rate Market
Juan Ayuso, Juan J Dolado and Simón Sosvilla-Rivero

Discussion Paper No. 627, February 1992 (IM)