Multi-commodity real options analysis of power plant investments: discounting endogenous risk structures

Abstract

The value of power generation technologies can be derived from the investment cost, the plant’s expected lifetime, and the discounted cash flows, the latter of which typically are a combination of several underlyings, such as the price of fuel, electricity, and CO\(_2\). To determine this value, most studies assume predefined, uniform, and constant discount rates, irrespective of the fact that the specific risk strongly varies with the technology concerned and also over time. In order to endogenize the technology-specific risk, we develop a new model that explicitly accounts for the (likewise technology-specific) combination of the underlyings. More specifically, we use a multivariate binomial tree real options approach for analyzing the value of different power plants (gas-fired and coal-fired, with and without carbon capture and storage (CCS); hydro; wind; photovoltaics) and for taking into account technical change. We further investigate the influence of alternative CO\(_2\) policies on the plants’ values, modeling the CO\(_2\) price in three different ways and for three different carbon price levels (5, 25, 45 €/t\(_\mathrm{CO_2}\)): (1) as a stochastic process (geometric Brownian motion), reflecting the price development in the Emissions Trading Scheme of the European Union (EU ETS); (2) as a (constrained) stochastic process with a price floor, and (3) as a deterministic carbon tax. From the model application, using data from German exchange-based markets and a much-cited pilot study on future energy strategies and scenarios in Germany, we find a strong preference for hard-coal power plants in the low CO\(_2\) price scenario (\(P_{2015}=\) €5) and a low value of waiting in case of using the real options model. For all price scenarios, the value of waiting is much higher for both the CO\(_2\) permits with a price floor and the CO\(_2\) tax policy. This leads, for the medium and high CO\(_2\) price scenario, to a dominance of the CCS power plants. In the high CO\(_2\) price scenario (\(P_{2015}=\) €45), the value of waiting only delays the investment decision in the case of the floored CO\(_2\) permit prices. For the two other policies, the model predicts an immediate investment in CCS power plants once the CCS technology becomes commercially available in 2020.