Cities should replace the current lengthy and uncertain permitting process with a simple system of fees. If tall heights create costs by blocking light or views, then form a reasonable estimate of those costs and charge the builder appropriately. If certain activities are noxious to neighbors, then we should estimate the social costs and charge builders for them, just as we should charge drivers for the costs of their congestion. Those taxes could then be given to the people who are suffering, such as the neighbors who lose light from a new construction project.Planetizen's response (I took the above quote from them as well) sums up the animosity against Coase: "But should developers be able to build whatever they want by giving a one-time payout to the neighbors? That sacrifices comprehensive, intentional planning for a one-time infusion of cash, and later generations may not thank us."
The fact that it sacrifices comprehensive planning is, of course, precisely Glaeser's point. There are generational issues to consider, and how you price the present value of perpetuities can be complicated, but the proposal shouldn't be dismissed outright. Just further proof that the Coase Theorem is an economic idea that's hard to popularize.