For complex dynamic systems, a modular control design process is often employed, wherein the overall design is partitioned into smaller modules. This paper considers a particular inner-loop/outer-loop modular control strategy in which the designer of the outer-loop module does not know the specifics of the inner loop but instead possesses a reference model that captures the ideal inner-loop input–output behavior. In the first part of this paper, we establish analytical properties of the modular reference-model-based design. In the second part, we introduce a novel mechanism, referred to as the modular control error compensation, which mitigates the performance loss that arises when the inner-loop reference model is not matched. We propose an iterative algorithm, using μ synthesis, to design this compensator to reduce performance loss on the basis of two concrete worst-case performance metrics. The effectiveness of the modular control strategy with the modular control error compensation is demonstrated through experimental results on an automotive system. Copyright © 2012 John Wiley & Sons, Ltd.