Skip to Main Content
Even though portfolio theory has increasingly been applied to analyze large-scale investments under uncertainty - and especially so in the electricity sector-most analysis so far has been based on the static mean-variance approach, which has two shortcomings: it fails to take into account irreversibility in the form of high sunk costs and the associated implications for optimal dynamic behavior. In addition, variance is not always the ideal risk measure, given that return or cost distributions are not necessarily normal. In fact, if losses are potentially large, a risk measure taking into account fat tails should be adopted. In this paper we generate distributions by optimizing investment behavior in a real options model, thus considering uncertainty and irreversibility at the plant level, and use them in a dynamic portfolio model, where the conditional value-at- risk (CVaR) is the risk measure. More specifically, we look at the dynamics of the optimal technology mix over a future time period conditional on the initial distribution of technologies. The application to investment in the electricity sector with uncertain climate change policy shows that this approach is not only useful from the aggregate investment point-of-view, but also for the purpose of assessing of policy effects.