Anmerkungen:
In: Revista de Economia del Rosario, Vol. 2, No. 1, 1999
Beschreibung:
The approach of structural decomposition of time series is used to estimate, by using the Kalman filter, the unobserved components of annual inflation in Colombia for the period 1989:12-1998:8. The evidence suggests that, within a univariate framework, the inflation has a stochastic level and a cycle that repeats, on average, every 16 months. With regard to forecasting, the best performance of the model is for more than six months ahead. Within a bivariate framework, the annual growth of M1 and inflation share the same cycle and the former Granger-causes the latter. The correlation of trend components of both variables is the highest when the trend of inflation leads the growth of M1 in18 months while the correlation of cyclical components of these variables is the highest when the cycle of M1 growth leads the cycle of inflation also in 18 months. May be this kind of associations could explain the difficulty to find the timing between inflation and growth of M1