Maybe it depends on what Stata version one uses. I work with Stata 13 and here I can neither find A nor u_t. However, finding a workaround is not to tricky. For example, to find the matrix A one can use Stata's built-in cholesky function: matrix sigma = e(Sigma)
matrix A = cholesky(sigma). To get the unorthogonalized shocks into your data frame you could predict the residuals for each equation and multiply them with the coefficients of the inverse matrix where you can obtain the inverse matrix with a built-in function again: matrix invA = inv(A).
↧
Re: Vector autoregressions in Stata
↧