The US government’s new consensus estimate of the social cost of carbon (SCC)—around $43 per ton of CO2 from a 2020 baseline—has met with some approval in academic and other circles (as we discussed yesterday). But some of the harshest criticisms, at least insofar as the blogosphere would interpret them, have come from MIT economist Robert Pindyck (short article and working paper). Pindyck’s critiques are important, though we do not agree with all the conclusions he draws from them—conclusions that have been misinterpreted by those opposed to climate policy based on the SCC.
Climate policy critics focus, not surprisingly, on the blunt title and first two words of the abstract of Pindyck’s working paper: Climate Change Policy, What Do the Models Tell Us….”Very little”. The models Pindyck references are integrated assessment models (IAMs), like Bill Nordhaus’ DICE model, on which the SCC is partly based. These models integrate a huge amount of information and science to estimate how much a change in CO2 emissions will affect global warming and the damages it causes. But Pindyck argues IAMs “have crucial flaws that make them close to useless as tools for policy analysis…[they] create a perception of knowledge and precision, but that perception is illusory and misleading.” This statement has been taken to imply that one should throw out the SCC estimates, as well as the models upon which they are based.
Critics, however, ignore Pindyck’s conclusion: “My criticism...of IAMs should not be taken to imply that because we know so little, nothing should be done about climate change right now. . . Quite the contrary.” In our belief, he then ends up basically (we hedge here because he says “some have argued” rather than “I argue”) endorsing the SCC estimate on precautionary grounds, imposing a carbon tax of that amount and revising this later as we learn more. Hardly a condemnation of the SCC! Indeed, a closer parsing of his statements suggests that his point is to stop overselling the precision of IAMs while pressing ahead with a SCC to get the process started. Beyond these broad points, is Pindyck providing something new here and, in particular, is he overselling the overselling of IAMs?
Pindyck starts by strongly criticizing IAMs for lack of any grounding in economic theory, calling them “completely ad hoc and of almost no predictive value, with estimates of economic loss amounting to pure guesswork.” While it is true that the underlying climatological assumptions, range of resultant physical outcomes, and their translation into changes in economic welfare are fragile, it seems to us a distortion to accuse IAMs of wholesale abandonment of reasoned economic logic and discipline.
First, while the models do simplistically render temperature change into GDP losses, they do so transparently. Further, they also incorporate results of highly detailed sector-specific damage estimates that themselves are based on economic reasoning. The simplicity of the GDP estimates is necessary given the worldwide and cross-sector scope of impacts. Further, Pindyck himself endorses the highly detailed studies embedded in these models, such as those addressing major crop losses in farm-intensive economies, energy availability constraints in industry, or threats to social welfare arising from real price increases in critical components of consumer expenditures. Thus, the simplicity argument seems oversold.
Pindyck’s second critique—that the models fail to tackle the possibility of catastrophic climate change—is not strictly correct. This area of study is arguably the most controversial from an economics and natural science perspective. In the interest of conservatism, the IAMs typically avoid any but a token bow to incorporating catastrophic impacts. So long as this is understood by SCC users (who are free to consider additional catastrophic impacts or different ways of considering such impacts when making policy judgments), this is not a problem.
The third critique concerns discount rates. Economists have long researched and fought over the appropriate rate for discounting future events, with no consensus emerging and lively debate continuing. Pindyck adds little new here, though he does give a valuable reminder of a critically important conceptual tool. As is common in government practice, the Interagency Working Group adopts a range of discount rates to show how sensitive results are to the rate adopted. The choice of particular discount rates used may be arguable, but the reasoning for these choices is transparent and no one knows the right answer. In any event, Pindyck does endorse a consensus process at the end of his piece, from which the choice of discount rates has arisen.
Thus, to us, this new critique of the IAMs, as well as the implied critique of the SCC, is much ado about nothing. Pindyck identifies real problems, but they are not new, and don’t justify inaction, as he himself argues.