I notice two points:
1. At least for gasoline, they are measuring the externalities of driving, not of gasoline. Bad news for EV drivers intent on saving the world on mile at a time, because most of the associated externalities are still present.
2. The estimated cost of carbon / GHG is small compared to the other external costs like accidents and congestion. This is a common result among economic analyses of carbon costs, and I often wonder about it. If you use a value associated with the marginal cost of abatement, I can see it being quite low. But that’s in the current context of nobody abating much anything. I wonder what it would be if you projected the marginal cost of 80% or 90% abatement. That is, if we were actually to solve the climate problem.
Or, another way of thinking about it: if GHG emissions are potentially going to make the earth uninhabitable, it seems like maybe they’re underestimating the external cost of carbon. Because there is limited cost data available for “the end of the world as we know it,” economists can be forgiven for working with the data they have but we, the careful reader should bear in mind the limits.