Using a standard cost/benefit approach, the policy recommendations generated by stipulating unacceptable "tipping points" appear very inefficient. For example, the current Waxman-Markey bill pending in Congress requires an 83-percent reduction in U.S. emissions (relative to the 2005 base level) by the year 2050. Yet, most models show that if the whole world were to adopt such an aggressive target, the costs would far outweigh the benefits. The latest calibration of Nordhaus's "DICE" model indicates that such strict cutbacks would yield more than $15 trillion in net costs, that is, costs net of benefits. The other models (which use CBAs) studied by the IPCC show similar results.
This is from Robert P. Murphy, "The Economics of Climate Change." In his article, Murphy takes as given that global warming is real, it is caused by humans, government policy can reduce global warming, and government can be trusted to do the efficient thing. He still concludes that the policies being proposed today are too extreme from a straight cost-benefit viewpoint. He then goes on to relax the last assumption about government, introducing some basic public choice.