Abstract:
In this thesis, we investigate the impact of transmission on the strategic behaviour of firms competing in deregulated electricity wholesale markets. Assuming uniform-price auctions and locational marginal pricing, we first investigate the properties of the dispatch problem over networks that are constrained and/or lossy. Without loops and losses we derive important results as to how price varies as a function of demand at each node. Whereas, for networks with loops and losses, we discuss the non-convexity of the dispatch problem and show that the optimal value function may be non-convex. We model the strategic firms as Cournot agents, and discuss how the assumptions surrounding the rationality of the agents can influence the equilibrium outcomes. Under a full-rationality assumption, we prove that over a lossless radial network, with firms owning single generators, the line capacities ensuring that a single-node Cournot equilibrium exists form a convex set. However, in the case of networks with loops or with firms owning multiple generators we provide counterexamples to this convexity. We investigate the prices coming from a two-node network with a lossy line and find the conditions such that a generator is guaranteed to have a quasi-concave revenue function. We use this result to prove that there exists a pure-strategy equilibrium for a Cournot game over the same network. In the final chapters we present some applications of this work. We first define a mixed-integer stochastic transmission planning model that uses the set of capacities derived earlier as constraints to ensure that the unconstrained equilibrium exists for all periods. Finally, we examine the effect of carbon charges over a two-node network and find that the imposition of a carbon charge may increase emissions when firms behave strategically over a constrained network.