Temporary and Permanent Components of Colombia's Output

Structural time series models, frequency domain analysis, the HP-filetr, and the Blanchard-Quah decomposition, are used to observe, some peculiarities of the business cycle. Such properties are those related the volatility of the temporary component and the duration of the business cycle during both 1925-1994 and 1950-1994. For the longer period we find that cycles between three and six years seem to be the most important for the variability of output; volatility is greater for GDP than for per capita GDP, except when the processes are linearly detrended. For period 1950-1994. GDP, except when the processes are linearly detrended. For period 1950-1994. Although the linear trend plus cycle model does not perform very well, cycles of about eight years seem to be most important for the cycle. The results of the Blanchard and Quah decomposition show that demand shocks have important explanatory attributes for output fluctuations. However, supply shocks, are dominant in the behaviour of output.