I am estimating the impact of a set of predictors (Xi) on both the mean and the deviation from the predicted mean of the dependent variable Y. The standard approach is to use Linear regression with multiplicative heteroscedasticity (e.g., regh or reghv command). However, my model for the mean equation is more complex (includes lagged dependent variable and a couple of endogenous predictors, and as such I am using xtabond to estimate it). Is there a more sophisticated version of regression with multiplicative heteroscedasticity (something like regh for dynamic panel data), or I need to estimate the regression for residuals manually?
Thank you!
Thank you!
