Long Horizon Uncovered Interest Parity Re-Assessed

We review the evidence for both short and long horizon uncovered interest parity (UIP) and rational expectations over the period up to 2011, extending the sample examined in Chinn and Meredith (2004) by nearly a decade. We find that the joint hypothesis of UIP and rational expectations (known as the unbiasedness hypothesis) holds better at long horizons than at short, although the effect is somewhat weaker than documented in Chinn and Meredith (2004). Using the formula for the slope coefficient, we identify potential sources for the difference in slope coefficients at different horizons. We attribute our weaker findings for long horizon unbiasedness for certain currencies partly to the advent of extraordinarily low interest rates associated with the zero interest bound in Japan and Switzerland.