This article examines twenty-two foreign exchange currencies vis-a-vis the Philippine peso for evidence of long-term dependence. This study uses three different methods: the classical rescaled range and the modified rescaled range of Andrew Lo (1991), and the rescaled variance statistic of Giraitis et al. (2003). This study also makes a comparative analysis using these three methods between the peso and twenty-two other foreign currencies. These empirical tests are conducted with daily data obtained during the period of January to December 2010. Using the modified rescaled range and the rescaled variance methods, we show that there is no significant long-range dependence in all examined currencies; however when using the classical rescaled range, we detect significant long-range dependence in many cases. The ARFIMA modeling also performed for those, showed long-term memory. The software-based tools compatible with MATLAB and Excel software are also developed to ease the arduous process of testing and finding the existence of long-range dependence in financial market time series datasets, based on modified rescaled range and rescaled variance statistics.
ISSN 0116-7111 (Print)
Philippine E-Journals
