A New Test for Market Efficiency and Uncovered Interest Parity

TL;DR

This paper introduces a new dynamic regression test for Uncovered Interest Parity that offers more efficient estimates and reveals stable, positive coefficients across currencies, challenging previous findings.

Contribution

It develops a novel single-equation dynamic regression method for UIP that improves efficiency and captures dynamic risk premia effects.

Findings

Coefficient estimates are positive and stable across currencies.

The new method outperforms traditional OLS with HAC in efficiency.

Results challenge previous studies showing negative coefficients.

Abstract

We suggest a new single-equation test for Uncovered Interest Parity (UIP) based on a dynamic regression approach. The method provides consistent and asymptotically efficient parameter estimates, and is not dependent on assumptions of strict exogeneity. This new approach is asymptotically more efficient than the common approach of using OLS with HAC robust standard errors in the static forward premium regression. The coefficient estimates when spot return changes are regressed on the forward premium are all positive and remarkably stable across currencies. These estimates are considerably larger than those of previous studies, which frequently find negative coefficients. The method also has the advantage of showing dynamic effects of risk premia, or other events that may lead to rejection of UIP or the efficient markets hypothesis.