Hello,
I am using time series data for testing the impact of infrastructure development on the human development. The predictors are highly correlated, which lead me to use the ridge regression to deal with the problem of multicollinearity. The second problem is the autocorrelation in the data. To deal with autocorrelation, I have taken lag value of dependent variable. The Durbin Watson test results seem ok after taking the lag of dependent variable. However, I am interested in using Newey-west SE correction along with Ridge regression, which not only deal with autocorrelation but also help in correcting heteroskedasticity. I am using command : ridgereg y x1 x2 x3 x4, model (grr2). and the Newey-west command is : newey y x1 x2 x3, lag(1) .
Now my question: is there any combined command to use the ridge regression and at the same time correcting the SE with newey-west ? Thank you.
I am using time series data for testing the impact of infrastructure development on the human development. The predictors are highly correlated, which lead me to use the ridge regression to deal with the problem of multicollinearity. The second problem is the autocorrelation in the data. To deal with autocorrelation, I have taken lag value of dependent variable. The Durbin Watson test results seem ok after taking the lag of dependent variable. However, I am interested in using Newey-west SE correction along with Ridge regression, which not only deal with autocorrelation but also help in correcting heteroskedasticity. I am using command : ridgereg y x1 x2 x3 x4, model (grr2). and the Newey-west command is : newey y x1 x2 x3, lag(1) .
Now my question: is there any combined command to use the ridge regression and at the same time correcting the SE with newey-west ? Thank you.
