Behaviour in the energy transition
How ABMs can improve in decision-support power for energy policies
In order to mitigate the undesirable effects of climate change, the energy transition is scaling up globally. To steer household decisions in favour of the energy transition, local and central Dutch governments often adopt behavioural design rules when they develop energy policies.
The use of behavioural insights in energy policies allows for the targeting of different sub-groups in the population and often pinpoints the most effective way of presenting an intervention. However, the system-level context will be overlooked if energy policies are designed with individual decision-making as the sole viewpoint.
Taking the (macro or meso) system into account in which (micro) decisions take place, is essential for two reasons. First, system structures can simply counteract individual behaviour (e.g. when complex market forces result in negative energy prices and households respond by using more, instead of less, energy). Second, system-level trends emerging from individual micro-decisions can be unpredictable due to the complex nature of the scaling process (e.g. when community resistance to an energy project can’t be predicted by the sum of the individual opinions of citizens).
Mariëlle Rietkerk
A technique well suited to exploring system-level effects is agent-based modelling (ABM). ABM excels in capturing macro-level phenomena emerging from micro-level decisions of individual agents like households. Agent-based models are the most insightful when they translate theories on individual decision-making and when the model is fed with formalizations that are based on empirical data from social science research methodologies (e.g. surveys, lab experiments and field experiments).
At present day, insights and research methodologies from behavioural science are only scarcely utilized in the design of ABMs. A lot of models lack behavioural realism, and the decision-support power for informing policies remains low. Also, perspectives on the relationship between psychology and the broader system remain underdeveloped.
The perception of hassle as a behavioural barrier to load shifting
This project (a PhD research) carries the title:
“Habits and hassles in the energy transition. Integrating behavioural science and energy system modelling to increase the decision support power of agent-based load shifting models to policymakers”.
In the Netherlands, the speed of the energy transition poses a threat to the energy security of households because the growing amount of renewable supply and demand of electricity puts pressure on the capacity of the electricity grid, with grid congestion as a result.
We, therefore, focus on a behavioural measure to reduce grid congestion: household load shifting. Load-shifting means that a practice of electricity consumption (e.g. doing the laundry or charging the electric vehicle, demanding a ‘load’ of electricity), is moved to a different time of the day when renewable electricity is available and grid congestion is low.
According to recent TU Delft research, a significant threat to household load shifting can be the perception of hassle (Hubert et al., 2024, see link). Hassles stand in the way of behavioural change because to overcome them, people need to add more effort to fulfil a certain task or to make a decision. Hassles can have a significant impact, acting as a behavioural barrier for energy behaviour (e.g. de Vries et al., 2019, see link and Ebrahimigharehbaghi et al. (2021, see link).
Research into hassle, and the hassle of household load shifting, is still scarce. There are no clearly defined behavioural theories that describe hassle factors, and empirical findings about hassle are somewhat hidden in behavioural literature (in the form of other concepts like friction, stress, effort, annoyance etc).
In order to elevate the decision support power of ABMs for policies that regard household load shifting, this project runs along four research lines.
Research line 1.
Covers the fit of a behavioural theory with the actual behaviour or decision-making of agents in an agent-based model. By comparing models that regard adoption behaviours (like buying solar panels) and habitual behaviours (like load shifting), this research line provides recommendations for improving the match between a used behavioural theory and the agents’ behaviour.
Research line 2.
Covers an analysis of internal and external factors underlying household load shifting behaviour, and the psychological theories that could explain load shifting. By teaching the technique of performing a behavioural analysis to modellers, this research line paves the way for a broad uptake of behavioural factors in load shifting ABMs.
Research line 3.
Covers the gathering of empirical data on the perception of hassle when households engage in load shifting. By conducting lab and field experiments, this research line reveals if/what hassle is perceived by Dutch households, and by which factors the perception of hassle is reinforced or reduced. It provides the empirical grounding for the inclusion of hassle factors in ABMs.
Research line 4.
Covers the application of hassle factors in ABMs. By implementing the aspect of hassle barriers into ABMs and reflecting on the differences in outcomes with and without hassle, this research line highlights the importance of including behavioural aspects in models.
With this research set-up, I aim to demonstrate how behavioural theory and empirical data from social science research can be translated into valuable input for agent-based models, in order to increase the decision-support power of these models to policymakers.