We propose a method for finding the cost function and state estimator to be used in MPC, in order to obtain the same controller as a given low-order controller when constraints are inactive, and hence when the MPC controller is operating within its baseline linear regime. This is a very useful starting point for the development of an MPC controller when a successful existing controller is known, but it is desired to add the constraint-handling capabilities of MPC, and perhaps other functionality such as fault-tolerance.