Revisiting the unit root hipothesis in macroeconomic series : a historical and empirical study for Argentina

Abstract

This paper is an application of different unit root tests to Argentina real GDP between 1810 and 2004. Particularly six tests are made: Augmented Dickey Fuller, Dickey Fuller GLS, Phillip-Perron, Kwiatkowski-Phillips-Schmidt-Shin, Elliott-Rothenberg-Stock Point- Optimal and Ng-Perron Test. The main finding is that all tests favor the existence of a unit root in Argentine GDP when including the whole time series between 1810 and 2004. When considering structural breaks, dividing the whole period in three sub-periods, whose limits follow the paper by Perron (1989), the non-stationarity of the GDP tend to be confirmed, although for some individual tests the evidence on non-stationarity becomes weaker or disappears, especially for the sub-period between the Great Depression and the Oil shock. As we will show, these results coincide with the most literature available for Argentina. The relevance of such behavior of the GDP through time is linked to the possible long-run consequences of policy shocks, particularly monetary (as suggested in a footnote by Nelson and Plosser (1982) about the theoretical possibility of a “Tobin effect”) and fiscal policies (as suggested by the traditional portion of the endogenous growth theory where public spending may alter the rate of growth of the steady state) which could not only affect the GDP in the short run but also in the long run.