BAYESIAN ANALYSIS OF THE UNIT ROOT IN REAL EXCHANGE RATES: THE NAFTA CASE
Abstract
The
Purchasing Power Parity (PPP) condition is often regarded as one of
the basic building blocks for several exchange rate models. Even when
the PPP condition is generally agreed to fail in the short run,
several studies seem to support it to some extent for long-run real
exchange rates.
Different
frictions in the markets seem to prevent the exchange rates from
behaving in the way predicted by the PPP condition. Many of these
market frictions, however, could be reduced by international trade
agreements. The North American Free Trade Agreement (NAFTA) is an
example of such agreements involving Canada, the United States, and
Mexico.
This work
explores the effect of the signature of the NAFTA in the Canadian and
Mexican real exchange rates from a Bayesian perspective. The Bayesian
approach to unit root testing has often been proposed as a reasonable
alternative especially in near non-stationary cases, a frequent
characteristic of real exchange rate series, where the classical
approach has been found to lack statistical power.
The
results, found in general to be in line with the common results
observed in the literature, seem to support the strengthening of the
series’ stationarity after the signature of the NAFTA, particularly
in the Canadian case. Nevertheless, new problems are introduced when
using the Bayesian approach that demand further development and
discussion.
Cuervo (2006)
