1 Modeling of Asymmetry between Gasoline and Crude oil prices: A Monte Carlo Comparison

An Engle-Granger two-step procedure is commonly used to estimate cointegrating vectors and consequently error-correction models. This paper uses Monte Carlo methods to demonstrate that the Engle-Granger two step method leads to biased estimates of asymmetric parameters and in some cases suggests symmetry in the estimated parameters. In contrast, the single equation error correction model (SEECM) and nonlinear least square (NLS) methods for simultaneous estimation of the cointegrating vector and the ECM, perform better in estimating the asymmetric parameters and making inferences on existence of asymmetry when the error terms in DGP are white noise. When the error terms in DGP are auto-correlated, NLS estimations are less biased and inferences are less likely to be misleading compared with SEECM and the Engle-Granger two-step procedures