Lucas Davis at the Energy Institute at Haas blog (emphasis added):
Nobody in recent years has done more to try to answer this question than Ian Parry, a University of Chicago trained economist now at the International Monetary Fund. Over the last 20 years Parry has written dozens of articles on this topic, including influential work on gasoline taxes, pollution externalities, traffic congestion, and the double dividend.
In his latest work, Parry has turned his attention in an ambitious new direction. Together with a team of IMF researchers, Parry set out to quantify energy externalities for 156 countries. The result is the aptly-named study, “Getting Energy Prices Right”. The summary can be accessed for free here, or the entire report can be purchased for $28 here. Following best practices for open science the team has also made all of the data from the project publicly-available here including all of the information on power plant locations, mortality rates, emissions factors and transport data.
These figures begin to give a sense of the scale of the project. You want to know local pollution damages from coal in Bangladesh? They have it. Traffic congestion damages from gasoline in Morocco? Sure. Vehicle accident externalities from diesel consumption in Sri Lanka? Yep. Previous studies had measured marginal damages for particular energy types and for particular individual countries, but this new work provides comprehensive estimates for the entire planet. ...
I’m a big believer in the idea that some number is better than no number. Quantifying the external costs from energy for 156 countries is an incredibly ambitious task that is impossible to get exactly right. Can the IMF numbers be refined and improved? Absolutely. But in striving to assign actual numbers to these externalities, Ian Parry and co-authors have taken a significant step forward. If policymakers are going to make informed decisions they need to be able to weigh the full social benefits and costs of different alternatives.
Some number is better than no number? Yes, always, at least in the context of environmental benefit-cost analysis (when the crucial step of sensitivity analysis is not ignored). But egads, that question gives me nightmares!