I. True/False/Uncertain - Briefly explain. No credit without an explanation (6 m
ID: 3049683 • Letter: I
Question
I. True/False/Uncertain - Briefly explain. No credit without an explanation (6 marks each) 1. More variation in the residuals improves the precision of estimation 2. The Almon Distributed Lag Model improves estimation by overcoming the degrees of freedom and serial correlation problems of the Distributed Lag Model. 3. The Lagrange Multiplier test for serial correlation is superior to the Durbin Watson d-test 1. The long run multiplier can be found by using non-stationary data. 5. The Error Correction Model requires unit root, non-stationary variables that are not cointegrated.Explanation / Answer
1. False.
Residual is the difference between the observed value and the predicted value of the dependent variable. The less variations in residual term there would be, the more precise the result be, because in that case, the observed value would be close to the predicted value.
2. False.
Common shortcoming of the distributed lag model is the lag lenght. Also a permanent change in one variable leading to temporary change in the other variable
3)
The answer to this is true.
The Durbin Watson test is a popular test for autocorrelation, but it has certain deficiencies in that it can only be applied for a first order autocorrelation process and also it does not allow lagged values of the dependant variable as one of the regressors in the regression. The Durbin Watson test also is applicable only for non stochastic regressors. The Lagrange Multiplier test does not have this limitation and helps to detect unplanned autocorrelation in a model. The Lagrange Multiplier test for serial autocorrelation is thus more verstaile than the Durbin Watson test. This statement is true.
4)
This statement is false.
The stationarity of the data is imperative for a time series process to be testable and for methods like ARIMA to be applied in the modelling process. The long run multiplier will be estimated by how much output changes as autonomous expenditure changes. So the time series needs to be stationery for the long run multiplier to be estimated. This statement is false.
5)
False
Error correction model requires the variable to be counteracted. They will not be covering and errors will not be corrected by simply and demand force