The use of an inner stabilising loop can be deployed to derive a generalised predictive control law with guaranteed stability, and can also be employed in the extension of this control law to the multivariable case. This approach affords useful insight into the mechanism that guarantees stability, but the development of this work is unnecessarily complicated. Recent work has overcome this problem in the scalar case, and in this note, the results are extended to derive a simple and efficient implementation of the stable generalised predictive control law for multivariable systems.