TERMS OF TRADE CYCLES IN EXTREME LAND ABUNDANT COUNTRIES, 1870-2009. SPECTRAL ANALYSIS.

Mar del Plata, Buenos Aires

Congreso; XLVI Reunión Anual de la AAEP; 2011

Spectral analysis is applied to the identification of terms of trade cycles in 1870-2009 for 5 land-abundant countries: Argentina, Australia, Canada, New Zealand, and Uruguay. Power density spectrum functions produce statistically significant spectrral peaks associated with long-run cycles ranging between 24 and 56 years. The variance decomposition shows that they account for a very substantial fraction, between 68 and 93$, of the TOT total variance. These results are very robust to changes in the choice of the truncation lag, as well as to the type of spectral window (Parzen and Bartlett) used in the estimations.