Modelling Power Plant Investment Behaviour: Three modular investment algorithms including hard and soft factors

Simulation models become increasingly important in energy policy analysis. A literature review showed that there is demand for analysing the effect of multiple more realistic investment algorithms on the outcomes of energy policy analysis. The reason is that energy policy analysis could be incomplete without insight in the implications of the assumed investment behaviour in simulation models. The research question is: How is the effectiveness of the EU-ETS mechanism affected by diverse investment algorithms in an electricity market simulation model? In EMlab-generation, three modular empirical data based algorithms are designed including behaviour with technology-preferences, credit-risk considerations and risk-averse behaviour. The results showed that more realistic investment behaviour culminates in all experiments in at least one or more technologies with substantially different investment patterns. Different investment patterns caused a lowered CO2 price volatility in most experiments in relation to homogeneous profit only behaviour indicating that the CO2 price might be a more stable investment signal than earlier assumed. The effectiveness of the EU-ETS mechanism remains for all experiments however doubtful due to the substantial CO2 price volatility of more than 100%. The necessity of stabilizing measures such as a price floor and/or ceiling proposed by previous studies is reinforced by the results. This research shows the importance of being aware of the implications of the assumed investment behaviour in simulation models used for energy policy analysis. Two recommendations to deal with the implications of the assumed investment behaviour in models is to design more flexible and modular investment algorithms. Flexibility and modularity in investment algorithms enable and support exploring the effect of different behavioural configurations on outcomes of energy policy analysis.