IDENTIFYING THE NATURE OF THE SEASONAL COMPONENT. APPLICATION FOR ROMANIA’S QUARTERLY EXPORTS BETWEEN 1990-2006

The purpose of this paper is to study the identification methods of the nature of the seasonal component of a time series. These methods are represented by the verifying tests of the unit root for the models of seasonal autoregressive processes: the HEGY test, the Franses test, etc. In practice, it has been demonstrated that the seasonal component is both deterministic and stochastic. The HEGY test allows identifying the nature of seasonal variations, but it is difficult to establish a limit between the two parts. The correct arbitration of the test’ results of seasonal autoregressive processes, with and without a deterministic component, makes it possible to choose the appropriate methods of seasonal variations elimination. This aspect is highlighted by analysing the time series defined by Romania’s quarterly exports between 1990 – 2006.