1 Introduction
The uncovered interest parity (UIP) condition, a cornerstone in assessing foreign exchange
market efficiency, has confronted a vast number of empirical tests which have produced
unfavorable results until recently.1 The UIP condition is a no-arbitrage condition between
investing in a domestic currency denominated asset and a foreign currency denominated
asset such that
Se
(1 + it,k) = (1 + it,k) -+ÿ- (1)
where it,k and i↑k are the interest rates on domestic and foreign currency denominated
assets with k periods to maturity respectively, St denotes the nominal exchange rate as
the domestic currency value of one unit of foreign currency at time t, and Ste+k denotes
the current market expectation of the nominal exchange rate of t + k given all available
information at t.
A proper documentation of the assumptions underlying this equation is crucial in in-
terpreting the empirical evidence on the UIP condition. These assumptions can be stated
as follows: investors are risk neutral; transaction costs are negligible; underlying assets are
identical in terms of liquidity, maturity and default risk; and there is a sufficient number of
investors with ample funds available for arbitrage.
Previous empirical literature reaches an estimable UIP condition by log-approximation
of equation (1) and imposing rational expectations:2
∆kst+k = it,k - it,k where ∆kst+k = st+k - st. (2)
The UIP condition, hence, can be assessed empirically through estimating
∆kst+k = β0 + β 1(it,k - it,k) + ut+k (3a)
and testing the joint hypothesis of β0 = 0, β1 = 1, and ut+k is orthogonal to the information
available at t. In addition, assuming the covered interest parity (CIP) condition, i.e. ftk -