*by **Scutaru, Cornelia**;
Saman, Corina and Stanica, Cristian*

Published in Romanian Journal of Economic Forecasting,
2009, volume 11 issue 3, 34-46

Requires a PDF viewer such as Xpdf
or Adobe
Acrobat Reader 332Kb |

We continue with the problem of the relation between predictability and complexity in the Romanian economy, analyzing other two components of GDP: domestic consumption and public consumption. The basic idea of this work is that the unpredictability of a system gives a measure of its complexity, so that in order to predict a future state of a complex system one must find the system structure explained by some simpler components that can be predicted. The complexity of the economic system is reflected in the synthetic macroeconomic indicators (GDP and its components). We find the principal components of the macroeconomic variables as a preprocessing step and model them as linear combinations of some simpler non observable predictable variables; we have constructed empirical models for domestic and public consumption; it was shown that these models are sufficiently accurate to predict for two or three periods ahead.

**Keywords**:
complexity, predictability, consumption, complexity pursuit, time series forecasting

**JEL Classification**:
E20, E27, C22, C51