You misunderstand my comment. To be useful for policy, the studies need to assess the _actual policy_. The reference you provided merely shows that new EVs are on average producing less GHG than new ICE cars. This is a separate question to whether this policy (banning new gas stations) will reduce GHG.
For example, you would also need to include opportunity costs of the policy in question (e.g. less energy is used to run an existing ICE vehicle than to manufacture new EVs) and estimates for the downsides (how much extra driving this will cause). Not to mention the chance that this policy does not affect EV adoption, and thus wastes resources better used building renewables (or any other good policy).
The report also has the downside that the data and code are not open access. Couple that with the fact that it comes from a political organization and I doubt the reliability of their analysis for policy. It's too easy to tweak assumptions to fit your preconceptions, and their assumptions are not published.
For example they mention some values as assuming 95% renewable penetration. They fail to mention how much additional GHG will need to be emitted to build and run grid scale storage which is required at such high % of renewables. There are huge differences in estimates of this cost from different institutions, so it is critical to know which values they used. The fact that they do not present these estimates as a range or confidence interval is a red flag that their analysis is biased.
I have read a lot of papers like this, your assumption that I haven't is unfair. I like EVs but I dislike the reflexive assumption from proponents that they are good in all circumstances and any policy supporting them is good policy.