Testing for Uncovered Interest Rate Parity Using Distributions Implied by FX Options

A popular test of market rationality and risk neutrality has been that of the uncovered interest rate parity hypothesis (UIP) using Fama regression. The test’s estimation strategy relies on a wide array of assumptions other than rationality and risk neutrality, such as large samples and the normality of expectational errors, which are unlikely to be met in the available samples. The regression results are therefore difficult to interpret and researchers appear reluctant to accept the failure of this test for UIP as a strong case for irrationality or risk aversion. We circumvent the binding restrictions of econometric regressions by suggesting that UIP be tested directly using the revealed information on distribution of future exchange rates implied by option prices. We design a series of tests based on implied risk-neutral distribution and tend not to reject the hypothesis.