We propose a new algorithm which allows easy estimation of Vector Autoregressions (VARs) featuring asymmetric priors and time varying volatilities, even when the cross sectional dimension of the system N is particularly large. The algorithm is based on a simple triangularisation which allows to...