David J.C. MacKay has a draft book out online, http://www.inference.phy.cam.ac.uk/mackay/ , which is worth a browse. He is due to give a talk here in a bit, which should prove interesting.
The book is an attempt to look at renewable energy from a broad-brush numbers point of view (in which I suspect it does a rather better jobs than “Heat” by Monbiot). But the bit I want to look at is the “offsetting” section, p143 onwards, the Story of Joan and Thambo. Joan flies in a plane, and offsets her emissions by helping Thambo keep warm by not burning his pile of coal. This seems fair enough: she emits X CO2 but her offset money allows Thambo to not emit X. The example assumes that all is above board, as it might not be in the real world. Yet for some reason MacKay seems to view this with suspicion, because the Joan+Thambo combination is not emitting *less* CO2. But this is silly… the purpose of the offset it only to remove what you emit (though you could buy 2* offsets; or just buy the offsets without travel).
What he doesn’t really address is what happens when we run out of such easy methods to offset CO2, which I suspect is a much bigger problem.
Incidentally, p148 has a nice graph of what happens versus the costs of a carbon tax. According to his estimates, sequestration of CO2 from thin air occurs at $130 (per tonne) and impact on US car-driving only occurs at $400. Carbon offsetting, though is available from $7.50.