IJSRP Logo
International Journal of Scientific and Research Publications

IJSRP, Volume 6, Issue 10, October 2016 Edition [ISSN 2250-3153]


Dynamic Time Series Regression: A Panacea for Spurious Correlations
      Emmanuel Alphonsus Akpan, Imoh Udo Moffat
Abstract: The study examined that the linear relationship between Gross Domestic Product (Y_t) and Money Supply (X_t) from 1981 to 2014 is spurious and could be avoided by dynamic regression modeling. The fact that spurious regression always results in misleading correlations between two time series was a big motivation for undertaking this study. Therefore, exploring data from the Central Bank of Nigeria Statistical Bulletin, we found that the linear relationship between the dependent variable (Y_t) and the independent variable (X_t) seemed spurious as the errors of the regression model were found to be autocorrelated. In a bid to correct this problem of spurious regression, we identified lags 0, -1 and -2 of X_t as predictors of Y_t using cross correlation function. Hence, the dynamic regression of current lag and past lags 1, 2 of X_t as predictors of Y_t revealed that the errors are uncorrelated and the coefficient of determination is as low as 0.2086, indicating that Y_t and X_t are totally unrelated.

Reference this Research Paper (copy & paste below code):

Emmanuel Alphonsus Akpan, Imoh Udo Moffat (2018); Dynamic Time Series Regression: A Panacea for Spurious Correlations; Int J Sci Res Publ 6(10) (ISSN: 2250-3153). http://www.ijsrp.org/research-paper-1016.php?rp=P585910
©️ Copyright 2011-2023 IJSRP - All rights reserved. Use of this web site signifies your agreement to the terms and conditions.