2 Uncovered Interest Parity and Developed Economies
Earlier empirical evidence for developed economies within the context of the UIP condition
is generally unfavorable. Specifically, the majority of the papers documented the so-called
“forward premium bias” (or forward discount anomaly), that is, the forward premium pre-
dicts the spot exchange rate movement in the wrong direction, i.e. β1 is negative in equation
(3b).4 This result is by and large robust to the estimation techniques and the data set, as
exposed in the surveys of Froot and Thaler (1990), Taylor (1995), Lewis (1995), Engel
(1996), Sarno (2005), Chinn (2006) and Isard (2006). Froot and Thaler (1990) report only
few studies where β1 in equation (3b) is positive and even in those, the estimates are less
than the hypothesized value of one. Lewis (1995) emphasizes that accounting for discrete
changes in the economic environment may improve the empirical results. Engel (1996) high-
lights the existence of time-varying risk premium as a source of forward premium puzzle,
and reports the limited success of the risk premium literature in explaining the forward
premium bias. Sarno (2005) reviews the literature that focuses on nonlinear dynamics of
deviations from the UIP condition and that incorporates term structure models. Chinn
(2006) analyzes the robustness of the results with respect to the time horizon.
Traditionally, departures from the UIP condition are attributed to non-rationality of
market expectations and/or risk aversion of agents that demand a premium for investing
in “risky” assets. Essentially, these two are tested jointly while estimating equation (3a)
or (3b), and a rejection of the UIP condition implies that rational expectations and/or risk
neutrality assumptions do not hold.
In addition to the aforementioned “text-book” reasons, it is possible to state other
reasons for the unfavorable empirical evidence for the UIP condition. These include exis-
tence of transaction costs, possible effects of central bank interventions, existence of limits
to speculation, and the possibility that investors may care for real rather than nominal
5
returns.5
We next analyze each of these potential sources of deviations from the UIP condition