The House Science Subcommittees on Environment and Oversight yesterday held a joint hearing titled “At What Cost? Examining the Social Cost of Carbon.” Four witnesses testified. I will review some hearing highlights, but first a bit of background.
The social cost of carbon (SCC) is a monetary estimate of the damages supposedly caused by an incremental ton of carbon dioxide emitted in a given year. Discernible in neither economic nor meteorological data, social cost values are guesstimates produced by computer programs called “integrated assessment models” (IAMs). What IAMs “integrate” is a speculative model of how carbon dioxide emissions will change the climate with a speculative model of how climate change will affect consumption, GDP, and health.
Social cost estimation is inherently speculative for two reasons. First, IAMs forecast climate change and the associated economic impacts over immense timespans vastly exceeding the limits of empirical validation. For example, to compute social cost of carbon values, the Obama administration’s Interagency Working Group (IWG) ran three models—known as FUND, DICE, and PAGE—out to the year 2300.
Second, modelers have wide latitude in choosing the inputs that determine model outputs. By fiddling with non-validated climate parameters, made-up damage functions, and discount rates, modelers can get just about any result they desire. What they seem to desire are big, scary numbers to justify costly regulations. Indeed, if they produce a social cost estimate high enough, modelers can make fossil fuels look unaffordable no matter how cheap, and renewables look like a bargain at any price.
And during the second Obama administration, social cost of carbon estimates rose dramatically. Values in the Interagency Working Group’s 2013 Technical Support Document were about 60 percent higher than those in the equivalent document in 2010—as if somehow in just four years climate change had become 60 percent worse, and carbon dioxide regulations 60 percent more valuable. Skeptics accused the administration was running a politically-motivated GIGO exercise.
Okay, on to the testimonies.
Ted Gayer of the Brookings Institution agreed with a common criticism of the Interagency Working Group. OMB Circular A-4, a directive from the White House Office of Management and Budget on how agencies should estimate regulatory costs and benefits, states:
Your analysis should focus on benefits and costs that accrue to citizens and residents of the United States. Where you choose to evaluate a regulation that is likely to have effects beyond the borders of the United States, these effects should be reported separately.
The IWG did exactly the reverse. In both the 2010 and 2013 Technical Support Documents, they reported the optional global social cost values but not the required domestic values. That matters because, thanks to America’s superior adaptive capabilities, any domestic damage from an incremental ton of carbon dioxide is bound to be much smaller than the global damage. By reporting only global but not domestic estimates, the IWG not only allows but requires agencies to dramatically inflate the perceived benefit-cost ratio of their regulations. Gayer explains:
The difference between global and domestic benefits of greenhouse gas regulations is significant, as the global measure is 4 to 14 times greater than the estimated domestic measure. For example, for its proposed regulations for existing power plants, the EPA estimated climate benefits amounting to $30 billion in 2030. However, the estimated domestic climate benefits only amount to $2-$7 billion, which is less than EPA’s estimated compliance costs for the rule of $7.3 billion. The use of a global social cost of carbon to estimate benefits means that agencies will adopt regulations that could cost Americans more than they receive in climate-related benefits. This approach could be especially problematic if U.S. actions simply shift emissions overseas.
Kevin Dayaratna of the Heritage Foundation explained how the government’s models are “extremely sensitive to very reasonable changes to assumptions,” allowing them to be “manipulated to produce a wide range of costs.” Examples:
- Reducing the models’ hubristic attempt to project climatic and economic interactions 300 years into the future to a somewhat less unrealistic time horizon of 150 years reduces some social cost estimates by as much as 25 percent.
- Using updated, empirically-based estimates of climate sensitivity (how much long-term warming results from a doubling of atmospheric carbon dioxide concentration) reduces the IWG’s cost estimates by up to 200 percent.
- Running the models with a 7 percent discount rate, as required by OMB Circular A-4 but omitted by the IWG, reduces the estimated cost by 75 percent compared to a 3 percent discount rate. Other things equal, the lower the rate used to discount the present value of future climate damages, the higher the cost.
- Under some reasonable alternative assumptions, the FUND model has a 70 percent probability of generating “negative SCC values,” meaning that carbon dioxide emissions produce net benefits.
Asked by Rep. Randy Weber (R-TX) why the IWG refused to use updated climate sensitivity estimates, Dayaratna opined that it’s because the results conflict with the administration’s agenda. With the updated analysis, IAMs generate “negative SCC values” (net carbon dioxide benefits) even when run with a 7 percent discount rate.
Using a clone of the Energy Information Administration’s (EIA) energy-market model, Dayaratna estimates that a carbon tax based on the IWG’s cost estimates would, by 2035, reduce U.S. employment by 400,000 jobs, decrease the cumulative income of a family of four by $20,000, increase electricity prices by 13-20 percent, and reduce aggregate GDP by $2.5 trillion. The benefit? By 2100, less than a 0.2°C reduction in global temperatures and less than a 2 cm reduction in sea levels.
Michael Greenstone of the University of Chicago argued that the social cost of carbon is “perhaps the most critical component” of climate policy regulations because it “allows for the calculation of the monetary benefits.” He disagreed with the criticism that the IWG should have included cost estimates using a 7 percent discount rate, offering several reasons why he thinks 7 percent is too high. The appropriateness of using any particular discount rate is certainly debatable, but beside the point.
Given agencies’ obvious incentive to skew social cost of carbon analysis to justify costly regulations, the IWG should have gone the extra mile to appear fair and balanced. Instead, the IWG flouted OMB Circular A-4. The Circular requires the use of both 3 percent and 7 percent discount rates. It also allows agencies to “consider a further sensitivity analysis” using a rate lower than 3 percent. But while the IWG availed itself of that option, and included values based on a 2.5 percent rate, it repeatedly refused to use a 7 percent rate. Dayaratna’s tables below reveal what the IWG was apparently trying to hide.
Greenstone knocks down a strawman when he claims critics believe the IWG should “only reflect damages that are projected to occur in the United States” despite climate change being a global issue of great concern to many countries. The real criticism, as noted above, is that the IWG published the optional global values but not the required—and much lower—domestic values.
Patrick Michaels of the Cato Institute criticized the scientific underpinnings of the IWG’s estimates. First, the IWG’s climate sensitivity assumptions are outdated and do not reflect more recent, empirically-based estimates, which are about 40 percent lower. Lower climate sensitivity not only means less warming and smaller climate impacts, but also less risk of catastrophic outcomes such as ice sheet collapse.
The chart below shows the range and best estimate of IWG’s sensitivity analysis, based on Row and Baker (2007), with the range and average of recent literature.
Michaels further noted that satellite and balloon-sensed bulk atmospheric temperatures over the past 38 years have warmed about half as much as projected by climate models. Those same errant models provide the sensitivity estimates relied on by Row and Baker and the IWG.
Although 2016 was the warmest year in the instrument record, much of that warming was due to a very strong El Niño, which has subsided. There has been little net warming since the previous big El Niño in 1998. The warming slowdown (“pause”) during the past 20 years is observable in both the University of Alabama Huntsville satellite record and the Hadley Center Climate Research Unit surface record. The data suggest the divergence between the modeled climate and the real climate will start to grow once again.
Michaels also discussed the IWG’s reliance on models that are structurally biased because they lack a significant carbon dioxide fertilization benefit. Based on crop yield data from carbon dioxide enrichment experiments and Food and Agriculture Organization economic data on 45 major food crops, Craig Idso estimates that carbon dioxide fertilization added $3.2 trillion to the value of global agricultural output since 1961. The FUND model also includes a significant, albeit much smaller, carbon dioxide benefit. However, the DICE and PAGE models have virtually no fertilization benefit. “Had the more comprehensive CO2 fertilization benefit identified by Idso (2013) been incorporated in all the IAMs, the three-model average SCC used by the IWG would have been greatly lowered, and likely even become negative in some IAM/discount rate combinations,” Michaels argued.