POLICY FORUM CLIMATE Bon Emissions to Global Economic Output) to 1.5% Per Year
Total Page:16
File Type:pdf, Size:1020Kb
POLICY FORUM CLIMATE bon emissions to global economic output) to 1.5% per year. Calibrating the DICE-99 model to alter- To Hedge or Not Against native climate sensitivities that span the range displayed in the figure was more in- volved, because the DICE model includes a an Uncertain Climate Future? parameter that reflects the inverse thermal capacity of the atmospheric layer and the 1* 2 2 Gary Yohe, Natasha Andronova, Michael Schlesinger upper oceans. Larger climate sensitivities were associated with smaller inverse ca- t has been over a decade since Nordhaus and the damages associated with green- pacity values, so that the model could (1) published his seminal paper on miti- house gas–induced temperature change match observed temperature data when run Igation policy for climate change. His (4). We assume that decision-makers eval- in the historical past. The parameter was question was “To slow or not to slow?”; his uate the economic merits of implementing defined from optimization of the global answer was derived from a traditional cost- near-term global mitigation policies start- temperature departures calculated by DICE benefit approach. He found that a tax levied ing in 2005 that will be in force for 30 and calibrated against the observed depar- on fossil fuel in proportion to its carbon years. They know that they will be able to tures from Jones and Moberg (6) for the content, which would climb over time at “correct” their policy in 2035, and we as- prescribed range of the climate sensitivities Downloaded from o roughly the rate of interest, maximized sume that decisions will be informed by from 1.5° to 9 C (7). global welfare. Although many more analy- perfect information about both the climate Modest near-term mitigation would ses of the same question have since been sensitivity and the policy target. Their goal maximize discounted GWP, even if no mit- published, his results are still robust if one will be to maximize the expected discount- igation was done after 2035 (see the sup- assumes a deterministic world in which de- ed value of gross world product (GWP, the porting online text). Achieving optimality cision-makers are prescient. However, no global equivalent of gross domestic prod- or even meeting specific concentration tar- http://science.sciencemag.org/ decision-maker has perfect foresight, and uct) across the range of options that will be gets would not, however, necessarily hold the uncertainty that clouds our view of the available at that time (see online material temperatures below the 2° to 3° range iden- future has led some to argue that near-term for details and definitions). tified by Smith and Hitz (8) and the mitigation of greenhouse gas emissions The uncertainty in our understanding of Intergovernmental Panel on Climate would be foolish. Such policy would impose the climate system against which these poli- Change (IPCC) (9), as a threshold above immediate costs, they argue, and have un- cies will be framed is portrayed in the figure which damages caused by gradual climate certain long-term benefits. (below). It shows a continuous cumulative change would climb dramatically, and by We take a different approach in this distribution function Schneider (10) and the Policy Forum by assuming that decision- (CDF) of climate sensi- 100 IPCC (9), as a thresh- makers will someday become so concerned tivity estimated by 90 old above which abrupt about the potential damages associated Andronova and Schles- 80 changes become much on February 11, 2018 70 with climate change that they will take ac- inger (5) (where cli- 60 more likely. We there- tion. Even though it is impossible to deter- mate sensitivity is the 50 fore focused our atten- 40 mine exactly what sort of mitigation target temperature increase 30 tion on mitigation these future policies might ultimately that results from a dou- function (%) 20 Andronova and Schlesinger (5) pathways designed to 10 This paper adopt, a “wait-and-see” approach may no bling of atmospheric density Cumulative limit temperature in- 0 longer be the best near-term policy choice. concentration of green- 0 2 4 6 8 10 creases to four targeted Should we move soon to intervene in glob- house gases relative to Climate sensitivity (°C) levels (recorded in the al energy markets as a hedge against the preindustrial levels). It Cumulative distributions of climate first row of the table, expected cost of meeting a currently un- also shows a version of sensitivity (5 ). next page) that straddle known policy target? the same CDF that al- this critical threshold. We follow the modeling approach lowed us, for reasons of practicality, to work We assumed that global policy-makers adopted in the hedging experiments con- with a limited number of sensitivities that would choose among these options in ducted by Manne (2) and Yohe (3) for the were nonetheless representative of the con- 2035, when the true climate sensitivity Energy Modeling Forum to explore the tinuous CDF. Each sensitivity is associated would be revealed; but each target was as- policy implications of extreme events. Our with a probability, so that it conformed with sumed to be equally likely for the purposes analysis is based on a modified version of the continuous version. Both representa- of setting near-term policy in 2005. DICE-99 (Dynamic Integrated Model of tions show that climate sensitivities as high Maximum discounted GWP was computed Climate and the Economy)—a widely re- as 9°C are possible. using the modified DICE-99 framework spected model of global economic activity Several structural and calibration modi- for initial 2005 taxes ranging from $0 to fications of the DICE-99 model were re- $50 per ton of carbon. Some combinations 1The Department of Economics, Wesleyan University, quired to accommodate the wide range dis- involved doing too little in the near term, Middletown, CT 06459, USA. 2The Climate Research played in the figure. Because responding to so GWP fell as downstream mitigation Group, Department of Atmospheric Sciences, high sensitivities could be expected to put “ramped-up” to achieve the prescribed University of Illinois at Urbana-Champaign, Urbana, enormous pressure on the consumption of temperature limit. Other combinations in- IL 61801, USA. fossil fuel, for example, we limited the rate volved doing too much in the near term, so *To whom correspondence should be addressed. E- at which the global economy could “decar- GWP again fell even though mitigation mail: [email protected] bonize” itself (i.e., reduce the ratio of car- could be “turned down” after 2035. An ini- 416 15 OCTOBER 2004 VOL 306 SCIENCE www.sciencemag.org Published by AAAS P OLICY FORUM tial tax of roughly $10 per ton of carbon table (11). The second column shows that a The specific cost estimates are, of (about 5¢ for a gallon of gasoline that 2° target could not be achieved, even if miti- course, highly dependent on the global would grow at the rate of interest over gation policy began in 2005, for climate sen- modeling context of the DICE-99 model, time) balances these two sources of loss to sitivities above 3°; they are “impossible the analytical decision to include only un- maximize expected GWP. now” in the parlance of the table. Second, 2° certainty about climate sensitivity in the Comparisons drawn from the DICE mod- and 2.5° targets could not be achieved if an analysis, and the identified boundaries of el across the requisite adjustments for the initial $10 tax policy were imposed in 2005 the “impossibility frontier”; i.e., the tem- $10 initial tax and for a wait-and-see policy for climate sensitivities above 4° and 6°, re- perature limits that could not be achieved in a “robustness” chart are displayed in the spectively (“impossible later” in the table). now and others that could not be achieved Doing nothing through 2035 would put 3° if mitigation were delayed for 30 years. In beyond the range of possibility if the climate addition, it is highly unlikely that many (if DISCOUNTED ADJUSTMENT COSTS sensitivity were 7° or higher. any) of the fundamental uncertainties asso- ($) GIVEN AN INITIAL TAX OF $10 An initial $10 tax policy is remarkably ciated with the climate problem will be re- Climate Temperature target robust across the remaining possibilities, as solved over the next 30 years. As a result, sensitivity (degrees) shown in the table. Discounted adjustment we should expect that “midcourse” correc- (degrees) 2 2.5 3 3.5 costs are smaller than $10 billion except tions will involve repeated hedging exercis- for high climate sensitivities near the bor- es and thus, relative to the modeling frame- 1.5 $0 $0 $0 $0 der of the impossibility frontier. A wait- work presented here, more uncertainty. The 2 $2 $1 $0 $0 and-see approach leaves the global econo- qualitative conclusion supporting modest 3 $4 $3 $1 $0 my open for far more serious adjustment near-term mitigation is, nonetheless, ex- 4 IL $6 $2 $0 costs. Except for higher targets with low tremely robust, because it is uncertainty Downloaded from sensitivities, doing nothing through 2035 that produces its value. Adding other 5 IL $12 $3 $0 imposes costs in excess of $20 billion in sources of uncertainty would simply add to 6 IN IL $4 $1 more than half of the possible cases and that value by widening the range of futures 7 IN IL $6 $3 significantly larger than $50 billion for low over which we must hedge. Uncertainty is 8 IN IL $9 $5 temperature targets even with lower cli- the reason for acting in the near term, and 9 IN IL $12 $9 mate sensitivities (12).