VOLUME XII
AUTUMN 2004

NON-LINEAL FORECASTING OF EXCHANGE RATES: AN APPLICATION OF THE GENETIC ALGORITHM
 
MARCOS ÁLVAREZ-DÍAZ
Department of Economics, Columbia University
ALBERTO ÁLVAREZ
Instituto Mediterráneo de Estudios Avanzados-IMEDEA (CSIC-UIB)
 
In this work we have applied a Genetic Algorithm to model and predict the weekly data on the Marc, Yen, Pound and Peseta exchange rates against the US Dollar. Our principal aim is to verify whether or not the use of a genetic algorithm improves on the results of a random walk model. The comparison is made considering both point predictions and the anticipation of either devaluations or appreciations.
 
Key words: exchange rates, non-linear forecasting, genetic algorithms.
JEL classification: C14, C53.

TO DOWNLOAD THIS PAPER