Last Friday on the Daily Kos Internet forum, an anonymous blogger called “Climate Denier Roundup” (CDR) posted a hit piece on a new study by Heritage Foundation statistician Kevin D. Dayaratna, Guelph University economics professor Ross McKitrick, and Competitive Enterprise Institute climate scientist Patrick J. Michaels. According to CDR, Dayaratna, McKitrick, and Michaels (DMM) claim “we should actually subsidize fossil fuel use because higher carbon dioxide levels are good for crops.” Their study says no such thing. DMM have no wish to play central planner or encourage others to do so. Neither are they “deniers,” as CDR frequently insinuates, beginning with the title of the post.
Before examining CDR’s critique of DMM’s study, some quick background on the main topic at issue may be helpful.
As DMM explain in the abstract of their study, “We explore the implications of recent empirical findings about CO2 fertilization and climate sensitivity on the social cost of carbon (SCC) in the FUND model.” What do those terms mean?
Carbon dioxide fertilization is the beneficial impact of rising atmospheric CO2 concentrations on plant growth, photosynthesis rates, and water-use efficiency.
Climate sensitivity is the amount of long-term warming produced by a doubling of atmospheric CO2 concentration.
The social cost of carbon is an estimate of the present value, in dollars, of the cumulative damages from an incremental ton of CO2 emitted in a particular year. That estimate also represents the monetary value of avoiding or reducing a ton of emissions in that year.
Federal agencies use three integrated assessment models (IAMs)—DICE, PAGE, and FUND—to estimate the social cost of carbon. The models are called “integrated” because they combine a climate model, which estimates the physical impacts of CO2 emissions, with an economic model, which estimates the dollar value of climate change impacts on consumption, agriculture, and other variables affecting human well-being.
Other things being equal:
- The higher the climate sensitivity assumed in an integrated assessment model, the higher the social cost of carbon, and vice versa.
- The higher the CO2 fertilization benefit assumed in a model, lower the social cost of carbon and vice versa.
- The lower the discount rate used when running a model, the higher the present value of future climate damages and emission reductions, and vice versa.
DMM note that DICE and PAGE assume the CO2 fertilization benefit is zero, which is not valid, so they set those models aside and focus on FUND which allows CO2 to benefit agriculture in some regions. But like DICE and PAGE, FUND was built long before most modern studies estimating CO2 effects on plant productivity were published, and before a series of recent studies estimating climate sensitivity based on historical climate data. They find that when the FUND model is run with updated empirical information on climate sensitivity and CO2 fertilization, the social cost of carbon drops to very small numbers and has about a 40 percent probability of even being negative through at least the mid-21st century. A negative cost is another way of saying a net benefit.
CDR claims that to project net CO2 benefits through 2050, DMM had to use “higher discount rates that downplay future suffering.” Not so. DMM used the same range of discount rates that other studies use, namely 2.5 to 7.0 percent, and they show the reader all the results. They also show that when the parameters are updated, the choice of discount rate no longer matters very much. For instance when the FUND model is run with Nicolas Lewis and Judith Curry’s climate sensitivity distribution and a 30 percent boost in agricultural productivity (beyond what is already in FUND) from doubled carbon dioxide concentration, the social cost of carbon remains negative “even at a 2.5 percent discount rate.” 2.5 percent was the lowest discount rate used by the Obama administration to estimate the social cost of carbon.
CDR claims “the authors also just totally write off the possibility of ‘future extreme or abrupt events,’ the sort of low-probability but highly-catastrophic events that are otherwise unimaginably expensive.” Again, this misrepresents the discussion of the issue in the DMM study (any reader comparing the actual paper to the CDR straw man version will realize how misleading the Daily Kos hit piece is). Like all users of the FUND and DICE models (including Obama’s EPA), DMM don’t model major catastrophes because they aren’t represented in those models. DMM note that if catastrophes were added in, the resulting SCC numbers would be higher, but also completely arbitrary because there is no data on which to estimate probabilities of major planetary catastrophes.
The PAGE model attempts to compute the present value of low-probability, high-impact events, such as collapse of the great ice sheets, but as the authors discuss, its results are very sensitive to small changes in the assumed probability of catastrophe. Also, PAGE has no carbon dioxide fertilization benefit function, hence cannot assess the social cost implications of ongoing CO2 fertilization research. Of the three models, only FUND is suitable for that purpose.
Does that mean PAGE and FUND are equally biased, just in different ways? No. It is far more problematic to ignore CO2 fertilization benefits documented by decades of laboratory and field experiments than to ignore speculative doomsday events that are universally regarded as highly unlikely and that may not occur for centuries or ever.
CDR complain that DMM assume “CO2 is a more potent fertilizer than the consensus recognizes.” The “consensus” does not assume the effect is zero (as DICE and PAGE do), nor does it still equal the FUND effect which was based on studies published in the mid-1990s. DMM draw on well-known studies published in the mainstream science literature to update the parameters. CDR provides no evidence the recent research is incorrect or that DMM misconstrue it.
CDR faults DMM for assuming “there will never be any abrupt or extreme events like we’re already seeing.” CDR’s meaning is unclear. If he has in mind negative but non-catastrophic effects from temperature variability or rapid warming in some regions, those are included in the FUND model and DMM take them into account.
In climate parlance, “abrupt” and “extreme” are not interchangeable terms. A hurricane is an extreme event but it is not an “abrupt” shift to a new climatic regime. If CDR is belaboring the low-probability doomsday events he invoked previously, the IPCC’s Fifth Assessment Report (AR5) concluded that during the 21st century, ice sheet collapse is “exceptionally unlikely,” Atlantic Ocean circulation shutdown is “very unlikely,” and catastrophic release of frozen methane is “very unlikely” (Table 12.4). The climate models underpinning AR5 had an average sensitivity of 3.2°C. As DMM point out, recent observationally constrained studies estimate climate sensitivities as low as 1.5°C. That DMM do not discuss the possibility of the world ending the day after tomorrow is eminently reasonable.
CDR tut-tuts that DMM make “very brave assumptions … given the stakes.” A little self-awareness would go a long way here. CDR makes brave assumptions about the nobility of his agenda. Policies aggressive enough to achieve carbon neutrality by 2050—the climate movement’s central objective—would have devastating impacts on energy prices and household incomes, hence on public health and welfare.
CDR suggests technology currently masks the harmfulness of climate change because yields would be falling absent industrial agriculture’s increasing use of fertilizers and genetically modified crops. Well, absent modern technology, most of us would starve and freeze in the dark. Climate risks, whether purely natural or partly anthropogenic, are relative to the adaptive capabilities human beings develop to survive and thrive. Increasing yields are evidence that human ingenuity aided by CO2 fertilization is more than a match for climate change, not that things are worse than they seem.
Similarly, CDR claims that “in the absence of climate change, yields may well have been much higher.” In terms of agricultural production, there’s zero evidence for this. As shown in DMM, since 1980—presumably the era of anthropogenic warming—there has been a gradual but remarkably stable exponential increase in yields that explains nearly 99 percent of year-to-year variability. That would not happen if climate change were progressively depressing production.
CDR concludes by parodying skeptical blogger Mike Adams, who in 2013 wrote: “If carbon dioxide is so bad for the planet, why do greenhouse growers buy CO2 generators to double plant growth?” CDR comments:
We’d like to suggest some sequels: “If floods are so bad for crops, why do farmers irrigate?” or maybe “If viruses are so bad, why do we put them in vaccinations?” Or what about “If pepper spray is so bad, why do we put pepper on food?” and “If we salt our food, why not salt our fields? Though knowing deniers, the real question is something more along the lines of “If eating lead paint chips is so bad, why do they taste so good?”
This putdown misfires in two ways. First, it uses false equivalents, because the beneficial item in the second half of each sentence is not the same as the harmful item it’s supposedly comparable to, whereas the CO2 referred to by Mike Adams is the same each way. CDR should have asked “If floods are so bad for crops, why do farmers drown their fields?” To which the answer is, they don’t. Likewise, doctors don’t use live infectious viruses in vaccines, we don’t put pepper spray on our food, and we don’t destroy our food with salt. The hit against Adams fails because the analogues are false.
Second, CDR evades the real questions raised by CO2 generators. If farmers pump CO2 into greenhouses, growth chambers, and open fields to increase the value of their produce, how can emitting CO2 be a purely negative externality? If CO2 generators double plant growth, how can social cost models legitimately omit CO2 fertilization benefits? Aren’t models that lack such benefits structurally biased and, thus, unfit for use in policy making?
In an op-ed published today in the Daily Signal, the Heritage Foundation’s Kevin Dayaratna concisely explains the common sense of the DMM study:
Yes, statistical models can be useful for understanding real-world phenomena. Any model, however, is only as a good as the assumptions from which it is composed. Improperly specified models can deceive the public, misguide policymakers, and result in big costs for ordinary Americans.
A point well understood by DMM, though seldom acknowledged by climate crusaders, is that even if integrated assessment models could definitively determine that the social cost of carbon is on the high end of the Obama administration’s estimates, we would still not know whether carbon pricing or regulation is prudent. To establish that, the SCC analysis would have to be paired with an equally rigorous assessment of the economic and social costs of carbon mitigation. The latter are substantial and include higher energy costs, slower GDP growth, lower household incomes, and the associated health and welfare losses. Worse, popular CO2 reduction targets may be impossible to meet without perpetuating deadly energy squalor in developing countries.
When presented with that objection, Obama officials replied that the benefits of carbon energy “are not relevant to the SCC itself,” which considers only the economic impacts of CO2 emissions. However, that just confirms our objection: SCC analysis looks at only one side of the ledger. It purports to quantify climate change risk while turning a blind eye to climate policy risk.
Even if CO2 emissions just enhanced the greenhouse effect without enhancing plant productivity, it would not necessarily follow that taxing, capping, or banning carbon-based energy serves the public interest. History makes that very clear. In the absence of motorized transport, farm machinery, fertilizers, pesticides, refrigeration, plastic packaging, and a host of other fossil fuel-supported technologies that have dramatically increased the productivity and efficiency of global food production, the human population over the past century and more would have been vastly smaller, hungrier, sicker, and shorter-lived.