On the 2nd of October, we had the pleasure to engage in a delightful presentation by visiting PhD student Martina Iori about her work on the retail electricity market and the collective switching behaviors caused by the market liberalization.
We learned that the electricity market of Italy was liberalized in July 2007, meaning that consumers may now choose an electricity supplier in a competitive market.
Martina is conducting an empirical study with two key questions in mind: first, what causes customers to switch between service providers, and second, what are the consequences of different switching behaviors?
She wisely split her analysis into 3 distinct levels: individual level, household level, and market level. Her data consists of a combination of a survey (42,494 individuals and 17,940 households) and geographical data which helps her localize and partition elements of the market into separate regions.
As a methodology, Martina chose a Bayesian modeling approach (logit and mixed logit) to model the probability of switching, given the empirical data. This method is especially interesting because even though Bayesian modeling is praised for its flexibility, consistency, and efficiency, it is not very common in the Network Science community. Her specific model aims at quantifying one very intuitive question: what is the probability of a given customer switching providers?
The research is ongoing but we already had a glance at some of the results. The question is simple: what are the variables that influence switching behavior? The results suggest a number of things. For instance, if a household had switched gas provider previously or if they had good access to generic information (like the Internet or other forms of media) they were more likely to switch. The factors negatively affecting switching (resulting in a low chance of switching) were high satisfaction with the service-specific information of the current provider, high economic resources, and old age. I think these results have intriguing socio-psychological implications, and I would definitely try finding causalities in that area. The results I found interesting from a more economic perspective were that the size of the households (number of members) positively affected the chance of switching, while high market concentration (having fewer options) had a negative effect.
All these results were crowned by a nice geographical map where the effects of the actual location were studied. This shows that going south gradually decreases the chance of switching while the north and Sardinia (which is an autonomous region) engage in more frequent switching.
Last but not least, Martina outlined that this work will form the baseline for an agent-based model, which will aim to test different policy implications during energy transition towards smart grids.
Overall I think we’ve seen the first developments of a compelling research project which can offer a great insight into the consequences of market liberalization and its effect on the economy and the collective behavior of the affected society.