Gauthier de Maere d'Aertrycke ENGIE CEEME Investment in generation capacity has traditionally been evaluated by computing the present value of future cash flows on the basis of a single discount rate for all equipment. The introduction of competition in the sector requires a more refined analysis: the risk exposure of a plant depends on its type and on the portfolio of other capacities of the system where it is operating. Also the overall risk exposure of the owner of the plant should be taken into account. Discount factors should thus be differentiated with respect to plants and possibly owners. Because investment in turn affects future risk exposures and hence discount factors, investment is subject to a circular reasoning that should be eliminated. The importance of risk on current power investment justifies a formal computable model of these effects. This is done in this paper by presenting stochastic equilibrium models of generation investment in incomplete financial markets. The models come in two versions depending on whether there exist financial products to partially hedge long term risk or not. Demand is fixed and the models involve producers and consumers. Both are risk averse and their behavior is represented by coherent risk functions. The sub-gradients of these functions can be interpreted as stochastic discount factors of the different agents for their different investments. Discount factors are thus endogenous. We use degree theory to prove existence and discuss multiplicity of solutions. We illustrate the model by interpreting results in terms of implicit discount rates inferred from the equilibrium. Joint work with Ibrahim Abada and Yves Smeers