Answer Happy

Consider the following model: Y, a+ BX, +u, where Y, and X, are integrated of order one a) When can we say the above regression is spurious? What does it mean? b) Under which conditions does cointegration between the variables hold? What does cointegration mean? c) Suppose the Engle-Granger test for cointegration and the Johansen Procedure for cointegration reveal the existence of cointegration between the two variables. However, the regression results given by the two approaches are different. Which approach is to be trusted and why?
Consider the following model: Y = a +BX, +u, where Y, and X, are integrated of order one a) When can we say the above regression is spurious? What does it mean? b) Under which conditions does cointegration between the variables hold? What does cointegration mean? c) Suppose the Engle-Granger test for cointegration and the Johansen Procedure for cointegration reveal the existence of cointegration between the two variables. However, the regression results given by the two approaches are different. Which approach is to be trusted and why?