Question

(1)Consider the following estimated model that relates the demand for money (Md) to interest rate (R), real income (Y) and the lagged money demand (Mt-1 )

                         ^

                 ln Mdt = 2.00 – 0.10 ln Rt + 0.70 ln Yt – 0.60 Mt-1  

                       Se                  (0.10)          (0.35)          (0.10)

                                  R squared = 0.90     DW = 1.80    n = 8

(a)Would it be appropriate to test for first order serial correlation in this model with the “regular” DW test ? Explain.

(b)Provide an alternative Durbin Watson test you think is suitable for this model and test for first order serial correlation at the 5% level of significance.

Answer 1

(a) No, because there is an assumption of the DW test which states that - In the regression model the dependent variable should not present as an explanatory variable lagged by one period.

Here the variable named Money Demand is used as the dependent variable and it is also used as an explanatory variable as well in form of lagged by one period. So the assumption of the DW test is violated here which does not allow us to perform the DW test here because it would be misleading.

Thank You.

(B) As a alternative of DW test one can use the " Ljung-Box Q Test ( sometimes called Portmanteau Test)" for testing serial correlation of first order.