Mr. Stephen Treacy:

On modelling for transport, it is a complex area and one that I touched on briefly earlier. Our modelling is primarily around energy demand. It is macroeconomic modelling. However, we do not just leave it there, as such; we work with others who do modelling for other reasons, such as Transport Infrastructure Ireland. We have also spoken to the NTA. It is something that we expect to do much of in the future.

The basic question the Chair asked was about a zero value for EVs. They do not produce any emissions in our projections for the transport sector, but the impact of their electricity demand is captured in the energy sector. That is modelled.

The Chair also asked about air quality. We model for air pollutant emissions. We use a system of emissions modelling called COPERT that factors in on-road and off-road emissions, including tyre and brake wear, which the Chair mentioned. That data is captured.

The Chair mentioned regional modelling, which is not something we do currently. Our modelling is done on a national basis. I do not have any real answers for him at the moment. EVs are most efficient where internal combustion engine vehicles are least efficient, which is in urban areas.

It is not necessarily clear-cut as to exactly where the greatest benefit will be achieved. It also depends on vehicle types and the size and weight of the vehicles that are being substituted. We are, therefore, working on many issues with the other actors in the transport sector that have other models, and particularly the NTA, which, from our discussions with it, is modelling the elements of the NPF where it can project forward the future employment and housing locations and future journeys that will be needed. We do engage with the other actors. This is to inform and adjust our model, which is primarily an energy model we take from the SEAI and Economic and Social Research Institute's way to remodel. Many of us are involved in it. As the Chair said, it does not represent the answer to all the questions. Other models have a better answer to some questions.