By Shawn Ritenour
When President Barack Obama wanted to curtail carbon dioxide emissions, he instructed his economic advisors to construct a way to calculate the emissions’ effect on society. The metric thus adopted by the EPA is called the “social cost of carbon” (SCC).
Right from the start we should note an important distinction: Carbon is an element; carbon dioxide a compound. Carbon is a solid; carbon dioxide a gas. Carbon, in the form of (for example) fly ash, dust, fine particulate matter, can harm health; carbon dioxide is harmless except at very high concentrations (above 10,000 parts per million), and even then only after long, uninterrupted exposure. Unlike carbon, carbon dioxide is odorless, colorless, and, except under conditions just described, nontoxic — indeed, indispensable to photosynthesis and thereby to all life. “Carbon” makes people think of black soot, smoke, smoggy skies; “carbon dioxide” doesn’t.
That’s why proponents of reducing carbon dioxide emissions call them “carbon emissions” instead. The term is deceptive and plays on ignorance and fear.
Now to the nitty gritty.
The EPA uses three models to estimate the SCC: Climate Framework for Uncertainty, Negotiation and Distribution (FUND); Dynamic Integrated Climate-Economy (DICE); and Policy Analysis of the Greenhouse Effect. Together such models are called integrated assessment models. While such attempts seem laudable, the models are nevertheless unhelpful at best and destructive at worst. All incorporate similar, grave weaknesses in how they attempt to calculate the SCC. For instance:
All include a host of assumptions and projections that are necessarily more ad hoc than based on empirical reality. In order to come up with a single numerical estimate of the social cost of emitting a metric ton of carbon dioxide, the models use projections of future emissions (which also requires projections of both future GDP growth and the amount of carbon dioxide emitted per dollar of GDP), future atmospheric carbon dioxide resulting from past, present, and future emissions, average global temperature changes resulting from higher carbon dioxide concentrations, and the economic impact, in terms of lost GDP and consumption, of higher global temperatures. They also include estimates of the cost of reducing carbon emissions and assumptions about the rate of social-time preference by which future money is discounted. That is a lot of projecting and guesstimating.
Understanding all of the assumptions and projections that go into these models helps one to understand why the results are so iffy as to be of no real help at all for policymakers. Case in point: In 2015, the EPA used one of its models to estimate that the SCC was $37 per metric ton. That same year another study published in Nature Climate Change estimated the cost at $220 a ton. That is a difference of almost six times! What gives? It turns out that these models are highly sensitive to the values of the parameters representing the projected effects of the variables mentioned above.
For example, Kevin Dayaratna, a Ph.D. statistician, and David Kreutzer, a Ph.D. economist, have co-authored a series of studies showing that using different rates of just one of the variables — social-time preference — has significant effects on the resulting estimates of the SCC generated by the DICE model. Holding other variables constant, a discount rate of 2.5% yields a social cost of a carbon of $87.69 while a discount rate of 7% yields a cost of $12.25. They also noted that using more up-to-date estimates of how sensitive climate temperature is to carbon dioxide reduced the estimate of the SCC by anywhere from 32 to 42 percent, depending on the magnitude of the discount rate used. And they showed that the FUND model is likewise very sensitive to small changes to model parameters. In fact, it even allows for large probabilities that the SCC could be negative, i.e., that increases in carbon dioxide emission could result in net benefit for society.
Potential positive effects of increased carbon dioxide emissions are often dismissed out of hand. Yet recent scientific literature indicates this might actually be the case. A 2016 study published in Nature Climate Change documents that a large part of the earth’s land featuring vegetation has benefited from significant greening over the last 35 years largely due to rising levels of atmospheric carbon dioxide. Other research documents increased crop yields under high carbon dioxide concentration, making food more abundant and affordable.
With climate enemies like this, who needs friends?
On top of all of this, these models calculate the SCC based on the above-mentioned projections for 300 years into the future. Yet who knows what energy technologies we’ll be using even 100 years from now? Could anyone in 1918 have predicted today’s technologies? People really have very little idea how their economic decisions made today will affect their situation three years from now, let alone 30 or 300. The idea that any economist can predict the quantitative effect of an action today on the economy 300 years from now would be laughable if not taken so seriously by politicians seeking excuses for policies to which they’re already committed on other grounds.
Wouldn’t it be nice if we could scientifically determine the cumulative costs or benefits that result over the next 300 years from our choices in the present? It may be nice, but it is impossible. Because these models produced such wildly different results depending on the projections and assumptions baked in the mathematical cake, economist Robert Pindyck concluded after an extensive review of such models that they are so badly flawed as to make them virtually useless for policy. Yet various environmentalists urge the EPA to use them anyway. Because of their grave, if not fatal, weaknesses we should be very circumspect indeed. And we should be especially wary of relying on them to justify policies that impoverish the least developed societies for the sake of who knows what.