A dynamic output feedback regulator design which cascades an observer with a dynamic state feedback control law is presented. A comparison is made with the usual regulator design in which an observer is cascaded with a proportional state feedback gain. It is shown that under certain conditions both approaches can lead to the same output feedback system. However, a comparison of the output feedback response with the responses of the two state feedback Bystems (dynamic state feedback and proportional state feedback) shows that the norm of the difference between the output feedback response and the set of dynamic state feedback responses is never greater than the norm of the difference between the output feedback response and the proportional feedback response. In practice this means that the set of possible responses with dynamic state feedback provides a good indication of the actual responses which will be achieved with output feedback (dynamic state feedback law plus observer). With the proposed design scheme the system transient response can therefore be approximately analysed using the state feedback system alone before the observer is designed. The control law and observer design problems ore therefore more completely decoupled than they are in the typical proportional gain plus observer approach.