[Editor Note: Robert Murphy’s 2009 views on the Andrew Weitzman “fat tails” argument for pricing carbon dioxide were presented yesterday. Today’s post shares Murphy’s review in 2016 of the same issue, part of his coauthored Policy Analysis (No. 801), “The Case Against a U.S. Carbon Tax,” with Patrick Michaels and Paul Knappenberger. ]
“Who would buy such an insurance policy?”
“Fat Tails” and Carbon Taxes as Insurance?
We note that the leaders in the pro-carbon tax camp are abandoning traditional cost-benefit analysis, claiming its use is inappropriate in the context of climate change. One reason given for this is concern over “fat tails”—concern that climate change could result in damages far greater than what is currently considered likely. Worries about fat tails lead some carbon tax proponents, like Harvard economist Martin Weitzman, to argue that, instead of treating a carbon tax as a policy response to a given (and known) negative externality, it should be considered a form of insurance pertaining to a catastrophe that might happen but with unknown likelihood.1
But that argument poses some serious problems. The most obvious is that the utility of such “insurance” is declining, given the emerging evidence that very large warming is unlikely. Beyond that, the whole purpose of the periodic IPCC reports was to produce a compilation of the consensus research to guide policymakers. But Weitzman and others argue that policymakers should be concerned about what we don’t know.2
That argument certainly has some merit, but, as economist David R. Henderson points out, broad-based uncertainty cuts both ways in the climate change policy debate. For example, it is possible that the Earth is headed into a period of prolonged cooling, in which case offsetting anthropogenic warming would be beneficial—meaning that a carbon tax would be undesirable.3 So why should one unlikely but troubling scenario shape our policy thinking but another unlikely but troubling scenario be ignored?
Another problem with Weitzman’s approach—as Nordhaus, among other critics, has pointed out4—is that it could be used to justify aggressive and costly policies against several low-probability catastrophic risks, including asteroid strikes, rogue artificial intelligence developments, and bioweapons. After all, we can’t rule out humanity’s destruction from a genetically engineered virus in the year 2100, and what’s worse we are not even sure how to construct the probability distribution on such events. Yet few people would argue that we should forfeit 5 percent of global output to reduce the likelihood of one of the latter improbable catastrophes. Why then do some people make that argument about climate change?
That question leads to another problem with the insurance analogy. With actual insurance, the risks are well known and quantifiable, and competition among insurers provides rates that are reasonable for the damages involved. Furthermore, for all practical purposes, buying insurance eliminates the (financial) risk.
Yet, to be analogous to the type of insurance that Weitzman and others are advocating, a homeowner would be told that a roving gang of arsonists might, decades from now, set his home on fire, that a fire policy would cost 5 percent of income every year until then, and that, even if the house were struck by the arsonists, the company would indemnify the owner for only some of the damages.
Who would buy such an insurance policy?
1. Martin L. Weitzman, “On Modeling and Interpreting the Economics of Catastrophic Climate Change,” Review of Economics and Statistics 91 (2009): 1-19,
3. David R. Henderson, “Uncertainty Can Go Both Ways,” Regulation 36 (2013): 50-51,
4. William Nordhaus, “An Analysis of the Dismal Theorem,” Cowles Foundation Discussion Paper No. 1686, January 20, 2009.
5. Bob Inglis and Arthur Laffer, “An Emissions Plan Conservatives Could Warm To,” New York Times, December 28, 2008,