New paper – Forecasting the electric vehicle market

At last!

The first paper stemming from my PhD thesis has been published by Transport Reviews, after a lengthy and exhaustive review process.

The paper started as my literature review, and it mainly focuses on methods to model the diffusion of electric vehicles, considering a rather annoying complication – in many cases they will replace an already existing car, and so basic diffusion models are not enough: they must be supplemented with models that consider substitution as well.

The paper mostly compares methods, considering their advantages and shortcomings. I guess the most fun part comes by the end, where we harvest forecast figures from the reviewed papers (where available) and compare them with actual market shares (again, where available). The main conclusions arrive at this point:

1. The forecasts closer to reality are, without exception, given by the pessimistic or “status quo” scenarios. As expected, electric vehicle diffusion if far from reaching its maximum potential, despite the enormous public investments.
2. The papers that achieve better forecasting results are those using mixed methods; i.e., a combination of simulation methods for both the consumer side (generally, agent-based models) and the supply/government side (typically, system dynamics models). These mixed models have the potentiality of accurately disentangling the complexity of the problem.
3. Regardless of the method, the quality of a forecast essentially depends on good input data and a well-grounded process of calibration and validation.

While the conclusions are not necessarily novel, they could be useful for modellers to consider when forecasting this market. Also there is still an overview of papers tackling this problem up until 2021, which might be useful for researchers.

The paper was co-written with my brilliant supervisor, Professor Elisabetta Cherchi. It has open access and is available via this link.