Abstract: There is a close link between the truncation lag and the unit root test power and size. Akaike Information Criterion (AIC) tend to select a small value of lag. The modified AIC (MAIC) may substantially improve the size. GLS detrending may improve the power. A good inference may arise from a conjunction between these two. This paper applied two Hong Kong annual GDP series for the 1970-2016 period. The standard augmented Dickey-Fuller test (ADF) and the Dickey-Fuller GLS (DF-GLS ERS) method were used. It suggests that MAIC may be crucial for the lag selection of DF-type tests.
Keywords: Information criterion, lag, power, size, unit root.
Title: Dissimilarities in Unit Root Computational Results Using Various Methods and Information Criteria
Author: Gaolu Zou
International Journal of Computer Science and Information Technology Research
ISSN 2348-1196 (print), ISSN 2348-120X (online)
Research Publish Journals
