Avaliação do poder de testes de raíz unitária em séries temporais AR(1) de diferentes comprimentos

Abstract

The study of time series is essential in various fields, such as economics and social sciences, as it aids in understanding and predicting phenomena over time. A crucial aspect is the stationarity of the series, which can be affected by unit roots, compromising the effectiveness of statistical methods. In this work, we focus on the comparative evaluation of the Augmented Dickey-Fuller (ADF), Phillips-Perron (PP), and Dickey-Fuller Generalized Least Squares (DF-GLS) unit root tests, using simulations on series of different lengths. The analysis aims to provide recommendations on the choice of tests given the specific conditions of each study. This effort contributes to enhancing the selection of appropriate analytical tools, thereby increasing the accuracy of research involving time series by investigating the impact of series length on the effectiveness of the tests.