On the efficiency of networked Stackelberg competition

We study the impact of strategic anticipative behavior in networked markets. We focus on the case of electricity markets and model the market as a game between a system operator (market maker) and generators at different nodes of the network. Generators submit quantity bids and the system operator balances demand and supply over the network subject to transmission constraints. We compare the efficiency of a networked Stackelberg equilibrium, where generators anticipate the market clearing actions of the market maker, with a networked Cournot equilibrium, where they do not. We show that networked Cournot equilibria always exists but its efficiency loss is unbounded in the worst case. In contrast, networked Stackelberg equilibria do not always exist, but in certain settings where they do exist, the efficiency loss may be bounded above by a constant.