Back in 2013, here at IER I discussed Robert Pindyck, a professor of economics and finance at MIT with decades of experience publishing books and articles on energy issues. Pindyck had just released a scathing journal article on the computer models being used to guide policymakers in the climate debate. Now, in a follow-up paper released last month, Pindyck elaborates on his critique of the current models, saying it is intellectually dishonest to continue using them. Furthermore, his suggestions on how to make climate policy without such computer simulations are even more illuminating on just how little the experts in this area know. In this post I will largely just quote from Pindyck’s paper to show the public the true state of affairs in what they have been told is a “settled science”—where this term is supposed to include a whole host of taxes and regulations.

Pindyck: Computer Models “Useless” and Their Use Is “Dishonest”

In his new paper, Pindyck first reiterates his problem with the Integrated Assessment Models (IAMs) that economists have constructed to take the estimates of global warming from the physical scientists, in order to produce estimates of economic damages from varying amounts of greenhouse gas emissions. These computer simulations are seemingly necessary in order to guide policymakers if they are going to issue regulations and/or enact carbon taxes to reduce emissions, because sensible policy requires a comparison of costs and benefits. Yet here is Pindyck’s opinion of such computer models of the global economy and climate:

…I would argue that calling these models “close to useless” is generous: IAM-based analyses of climate policy create a perception of knowledge and precision that is illusory, and can fool policy-makers into thinking that the forecasts the models generate have some kind of scientific legitimacy. IAMs can be misleading – and are inappropriate – as guides for policy, and yet they have been used by the government to estimate the social cost of carbon (SCC) and evaluate tax and abatement policies. [Pindyck 2015, bold added.]

One of Pindyck’s problems with these Integrated Assessment Models, or IAMs, is that they are so flexible that the researcher can get out any desired answer. For example, holding every other setting constant, the “social cost of carbon” can vary from $11/ton to more than $200/ton simply by varying the discount rate used. If we use the market rate of return to capital, we get a really low figure, whereas if we push the discount rate down close to zero, then we get the very high numbers. The choice of a discount rate has nothing at all to do with the so-called “climate consensus”—meaning the physical science of climate; physicists and chemists can’t tell government officials how to compare $1000 in economic damages today from a carbon tax, against $1000 in avoided climate change damages a century from now.

Another huge problem that most people would be shocked to learn is that there is neither theory nor data to back up the way these computer models relate a specified temperature increase to a predicted amount of damage. As Pindyck explains it:

One of the most important parts of an IAM is the damage function, i.e., the relationship between an increase in temperature and GDP (or the growth rate of GDP). When assessing [the climate’s sensitivity to emissions], we can at least draw on the underlying physical science and argue coherently about the relevant probability distributions. But when it comes to the damage function, we know virtually nothing – there is no theory and no data that we can draw from. As a result, developers of IAMs simply make up arbitrary functional forms and corresponding parameter values. [Pindyck 2015, bold added, footnotes removed.]

Thus we see that all of the fancy computer models—including the three that the Obama Administration Working Group selected to estimate the “social cost of carbon”—rest on quicksand.[i] Most policymakers, let alone the general public, have no idea how flimsy and arbitrary is the foundation upon which these computer simulations stand. This is what leads Pindyck to write: “I will argue that the use of IAMs to estimate the SCC [social cost of carbon] or evaluate alternative policies is in some ways dishonest, in that it creates a veneer of scientific legitimacy that is misleading.” Later in his paper Pindyck further writes that “the developers and users of IAMs have tended to oversell their validity, and have failed to be clear about their inadequacies.” Because of this overselling of the power of these models, Pindyck believes “[t]he result is that policy makers who rely on the projections of IAMs are being misled.”

Pindyck Admits That Unlikely Catastrophes Are the Only Game In Town

Now it would be one thing if these scathing remarks came from a young staffer at a right-wing think tank that opposed carbon taxes. Yet to repeat, Pindyck is a veteran in energy economics who teaches at MIT, and he supports government action to mitigate climate change. Thus he has no vested interest in being so frank about the uselessness and misleading nature of the computer models generating estimates of the monetary damages from carbon dioxide emissions. Pindyck, to his credit, is a straight shooter and is letting everybody know what the real situation is.

What is even more revealing, however, is to see what Pindyck thinks a better approach would be. After all, if we throw out the existing economic simulations, how can policymakers know how high to set a carbon tax? Here’s part of Pindyck’s answer:

This does not mean we have to throw up our hands and give up on the estimation of the SCC and the analysis of climate change policy more generally. I have argued that the problem is somewhat simplified by the fact that what matters for policy is the possibility of a catastrophic climate outcome. How probable is such an outcome (or set of outcomes), and how bad would they be? And by how much would emissions have to be reduced to avoid these outcomes? I have argued that the best we can do at this point is come up with plausible answers to these questions… [Pindyck 2015, bold added.]

We have already left the (utterly arbitrary and misleading) world of smooth functions which calculate the marginal impact of a further ton of emissions, allowing policymakers to set the “optimal” carbon tax. Pindyck spends most of his paper explaining why that is a delusion.

Instead, as the block quotation above indicates, Pindyck simply wants experts to pick a few catastrophic scenarios, giving estimates of their severity and the associated levels of emissions making them more or less likely. This would give policymakers a rough idea of what they would need to do, in order to achieve reductions in the likelihood of such catastrophes.

To understand the enormity of the chasm between the current approach and what Pindyck is suggesting, we should follow up on the phrase I bolded in the quotation above. Specifically, why did Pindyck write that “what matters for policy” are just the catastrophic scenarios? Earlier in his paper he had explained:

How do we know that the possibility of a catastrophic outcome is what matters for the SCC? Because unless we are ready to accept a discount rate that is very small, the “most likely” scenarios for climate change simply don’t generate enough damages – in present value terms – to matter. [Pindyck 2015, bold added.]

At this point, I want the reader to pause, take a breath, and grasp the bombshell that Pindyck just lobbed. He is confirming what I have been telling IER readers for years: Using the very computer models and estimates from the IPCC’s own reports, I can show that the standard UN climate targets (such as limiting warming to 2 degrees Celsius) do not pass a standard cost/benefit test. Notwithstanding all of the rhetoric about the “science is settled” and how we are (supposedly) seeing the awful consequences of human-caused climate change before our very eyes, the IPCC’s own reports show that the popular “solutions” to the problem of climate change don’t make any sense.

Once we realize this awkward fact, everything else falls into place. It’s why climate activists are trying to discredit the use of GDP as a criterion for policy evaluation, and it’s why the IPCC itself is switching its rhetoric to risk management and a focus on “co-benefits.” They have to do these things, because they know their own reports better than the outside world, and they know full well that when using standard economic tools, even mildly aggressive climate policy targets cannot be justified.


Robert Pindyck believes that the government ought to take steps to mitigate the possibility of catastrophic climate change. Pindyck is intellectually honest, however, and so he is quite frank about the useless and arbitrary computer models that are currently being used to mislead both policymakers and the public, especially with the false precision of their estimates of the “social cost of carbon.” The people citing these “official” estimates of such-and-such dollars per ton, as if they were analogous to measurements of the sun’s mass, are either being deceptive or have themselves been duped.

Far from being optimally calibrated using an analysis of marginal costs and benefits—the way most economists describe it to the public—Pindyck openly admits that if the government wants to justify aggressive action against greenhouse gas emissions, it’s going to rely on a small group of experts simply making guesses about what should be done, in order to reduce the probability of vaguely defined catastrophes that even the experts admit probably won’t happen if governments do nothing. The case for aggressive government intervention keeps getting weaker and weaker, and yet the rhetoric against “deniers” continues to ratchet upward.

[i] Actually the case is even worse than Pindyck suggests. It’s not really the case, as he claims, that we have no data on the connection between temperature increases and the effect on GDP. We know that the modest increase in temperature since preindustrial times has gone hand-in-hand with, well, the Industrial Revolution. Most people today would be glad to take the warming plus the GDP growth that humanity has experienced since 1750, if forced to choose as a package deal. Obviously that consideration doesn’t prove that history will repeat itself going forward, but my modest point is that, contrary to Pindyck’s claim, we actually do have some empirical evidence regarding the effect of global warming on human welfare, and thus far the evidence says it is benign. To get catastrophic projections, we have to build computer simulations in which things unfold differently from how they have done so according to our observations.

Print Friendly, PDF & Email