As tactical military networks deploy new IP-capable radios (e.g., JTRS), it is expected that network connectivity will increase both within and among Service and Coalition networking domains (e.g., Army, Navy, NATO). Joint commanders will likely desire control of this increased connectivity to complete their mission goals. One mechanism known to (coarsely) manage multiple disparate IP networks is BGP routing policy. As such, this paper presents an experimental framework to estimate BGP's policy impacts on mission outcomes (e.g., expected Blue/Red casualties). In support of this effort, a testing platform was developed by integrating an agent-based combat simulation tool with a large-scale network emulation platform. Here, the simulation tool provides military mission modeling (e.g., force movements) and metrics (e.g., casualties counts) while the network emulation platform models the mission's underlying communications network (e.g., link characteristics, protocols, policy). Using this testbed, a sample military mission was evaluated over three communication architectures. The first architecture connected (Blue) mission units together using (fixed) point-to-point circuits, whereas the latter two used shortest-path (dynamic) routing and policy-based (dynamic) routing, respectively. It was found that both dynamically routed network architectures provided better Blue/Red casualty ratios and higher network reachability than the fixed, point-to-point network architecture. However, the addition of BGP policy did degrade performance relative to short-path routing. Furthermore, the degradation in mission and network performance did vary based on the BGP policies implemented. This paper's results imply that if policy-based routing is to be considered as a network management element for newly deployed joint IP networks, further experimental assessment is needed to understand the inter-relationships between BGP policy, mission outcomes, and network performance.