ARDL model as a remedy for spurious regression: problems, performance and prospectus

Spurious regression have performed a vital role in the construction of contemporary time series econometrics and have developed many tools employed in applied macroeconomics. The conventional Econometrics has limitations in the treatment of spurious regression in non-stationary time series. While reviewing a well-established study of Granger and Newbold (1974) we realized that the experiments constituted in this paper lacked Lag Dynamics thus leading to spurious regression. As a result of this paper, in conventional Econometrics, the Unit root and Cointegration analysis have become the only ways to circumvent the spurious regression. These procedures are also equally capricious because of some specification decisions like, choice of the deterministic part, structural breaks, autoregressive lag length choice and innovation process distribution. This study explores an alternative treatment for spurious regression. We concluded that it is the missing variable (lag values) that are the major cause of spurious regression therefore an alternative way to look at the problem of spurious regression takes us back to the missing variable which further leads to ARDL Model. The study mainly focus on Monte Carlo simulations. The results are providing justification, that ARDL model can be used as an alternative tool to avoid the spurious regression problem.