Capex temporal allocation and cost recovery in long-term generation capacity expansion planning

Abstract

This work revisits some interesting properties of long-term generation models and sheds light on cost-recovery properties associated to this kind of representations in the framework of marginalist remuneration electricity markets. We propose a sound way of dealing with the allocation of investment cost (CapEx) when analyzing minimum-cost generation capacity expansion planning by using overall investment cost. We compare it with the most common approach based on cost annualization that usually disregards or uses a simplistic representation of future years. Overall-cost-based representation shows some advantages with respect to the annualized alternative. First, it does not require fixing a criterion for investment cost annualization, second, it provides a coherent definition of long-run cost (or prices) when assuming infinite life companies by modelling residual values, third, it allows for an adequate representation of the life span of generation assets and finally, it allows for a yearly tracking of company cost what it is very useful if taxes or financial constraints are to be detailed included in the model. On the other hand, the cost recovery property in marginal pricing, which holds for annualized cost representation under some reasonable hypothesis, it is not satisfied for overall cost representations in some situations. We analyse in detail, some conditions that must be fulfilled to get full cost recovery, showing a connection between cost recovery and the presence of investments in the last year of the studied horizon. Besides, some case studies are shown to check the effect of overall cost representation in scenarios with demand shape modifications, inclusion of new technologies or closure of old plants. Additionally, the relationship between long- and short-run costs (or prices) in this context is analysed, showing that, due to a duality gap in the model, some differences appear, that can be reduced by decreasing the size of the time periods of the study.