Your study is based on a land value of $2 million per acre. My first Google result shows that Manhattan is, at its most expensive, $90 million per acre: https:

Your study is based on a land value of $2 million per acre. My first Google result shows that Manhattan is, at its most expensive, $90 million per acre: https://cityroom.blogs.nytimes.com/2008/05/13/2100-a-square-foot-for-vacant-land/ https://cityroom.blogs.nytimes.com/2008/05/13/2100-a-square-... This matches my intuition that cars are much more massively subsidized in city centers (where they should be heavily discouraged) than in suburbs (where they're mostly fine). I also think roads have way higher value than other land, because they connect spaces. Think of it this way: you can build 99% of a wall, and I can still cross from one side to the other, but if you build 100% of a wall, I can't. So that last 1% of a wall has significantly higher externality cost than the first 99% of a wall. Similarly, the existence of a road makes it impossible to cross a street without waiting. It's the last 1% of the land necessary for going from one part of a city to another, and so it's way more valuable than the price per acre would suggest. Your study also doesn't include infrastructure costs, and also costs in inconvenience to pedestrians. Your study also estimates congestion cost based on the assumption that 10% of travel is under congested conditions, but in Manhattan this is closer to 100%. All your study's costs are conservative estimates. In conclusion, I completely agree with the overall conclusion that car usage in most of the US has relatively low externality costs. I just massively disagree that car usage in a superdense city center like Manhattan also has relatively low externality costs.