In a blog post a while back, I pointed out that William Nordhaus's work on the economics of global warming demonstrated a serious problem with policy arguments based on externalities—the risk, in dealing with something as uncertain as costs and benefits over the next century, that the calculation of externalities will be biased in the direction that produces the result the person doing the calculation wants. Nordhaus included, as an essential part of his argument, an estimate of the expected cost from low probability/high cost outcomes of permitting global warming but made no attempt to include the cost of equivalent outcomes from preventing it—although there is no strong reason to assume that the current global climate is optimal, and so that any change will produce costs but not benefits.
Robert Murphy has webbed a detailed response to a recent Op-Ed by Nordhaus on global warming, dealing with a variety of other issues. Among his points:
1. Climate models have, on average, overpredicted warming, suggesting that a better estimate would be about 2/3 the current predictions of such models.
2. The paper which Nordhaus cites as supporting the claim that CO2 is a pollutant actually provides evidence that net externalities are positive over a temperature range representing about the next fifty or sixty years of predicted rise, and go negative only after that.
3. Nordhaus's own work finds that the optimal policy to reduce global warming, coordinated among all nations of the world, produces net benefits, but that some proposed policies, including one proposed by Gore, produce much larger net costs.
One point Murphy did not make but that is worth noting is that the benefits of climate control, on Nordhaus's own figures, are not very large. The optimal policy—for obvious reasons not likely to occur—is calculated to produce a net benefit of about three trillion dollars. That sounds like a lot of money—until one recognizes that it is spread over the entire world and about ninety years. That makes the annual benefit of the ideal policy about 33 billion dollar a year—roughly one percent of the current U.S. federal budget or one tenth of a percent of current world income.
Which suggests that, with a less ideal and more realistic policy, net costs are likely to be larger than net benefits.