Instead of using the usual calculations for the variance components of our model, we use a method that is robust to heteroscedasticity and serial correlation. The usual estimate for the variance-covariance matrix is as follows:
Instead of using the usual calculations for the variance components of our model, we use a method that is robust to heteroscedasticity and serial correlation. The usual estimate for the variance-covariance matrix is as follows:
3.138.120.136