Cognitive radios permit dynamic control of physical layer resources such as transmission power and constellation size; these degrees of freedom can be employed to achieve significant improvements in network throughput above that obtainable using conventional radios (with fixed transmission power and constellation size). In this paper we present a unified framework for coordinated resource allocation across the entire protocol stack: physical, medium access, network, and transport layers. Our focus is on demonstrating that joint optimization over transmission power, constellation size, scheduling, and multicommodity flow can achieve greater network performance over optimizing resource allocation at each layer individually. We present three cases where a modularized network design problem can be "merged" and then characterize the benefit achieved by the merger.