1. After Copenhagen & Cancun: What is the Politically Feasible Path to Comprehensive Targets for Greenhouse Gas Emissions? Jeffrey Frankel Harpel Professor, Harvard Kennedy School Harpel Professor, Harvard Kennedy School The Marvin & Carolyn Birger Lecture, Tufts University, March 16, 2011

2. 2 The recent Cancun Conference of the Parties to the UN Framework Convention on Climate Change (Dec. 2010) was considered a success, –in part because expectations had been so low. The preceding year’s Copenhagen meeting (Dec. 2009) was considered a failure, –in part because expectations had been so high.

3. 3 What is the definition of progress? It is useless to evaluate negotiations by whether they produce a sweeping communiqué, –such as Bali or G8 agreement to limit warming to 2°C at L’Aquila, Italy, in 2009: They are just words. –Modest targets, if short-term & credible, are worth far more than aggressive targets, if long-term & non-credible.

4. 4 Definition of progress? Keep in mind the Herculean tasks of bridging the gap between rich countries & poor, the gap between environmental aspirations & economic costs that people are willing to pay, the gap between what leaders say, & what commitments are enforceable/credible. – Progress ≡ steps toward specific credible commitments by a large number of countries.

5. 5 Assume that the climate regime will continue to follow the current path Features of the Kyoto Protocol worth building on -- –Politics: Quantitative limits maximize national sovereignty –Economics: Market mechanisms, esp. international permit trading –Thus (2001) “You’re Getting Warmer: The Most Feasible Path for Addressing Global Climate Change Does Run Through Kyoto.” What was sorely missing from Kyoto: –Participation by US, China, & other developing countries –A mechanism for setting targets further into the future, past 2012 –Any reason to expect compliance.

6. 6 Progress of a sort Most countries (80) responded to the Copenhagen Accord in 2010 by submitting plans for reducing emissions. By the time of Cancun, 21 countries had associated themselves with specific quantitative targets counting the EU27 as one and including 7 big non-Annex-I countries. Of course some, like China or US, are vague about seriousness of commitment. Also India & China’s 2020 target ≈ BAU (Business as Usual). But that is not a problem.

7. 7 unlike other approaches based purely on: –Science (concentration goals), –Ethics (equal emission rights per capita), –or Economics (cost-benefit optimization). Why the political approach? –Countries will not accept burdens they view as unfair. –Above certain thresholds for economic costs, they will drop out. My Proposal: formulas for pragmatic targets, based on what emission paths are possible politically:

8. 8 “An Elaborated Proposal For Global Climate Policy Architecture: Specific Formulas and Emission Targets for All Countries in All Decades” (2009) suggested a framework of formulas that produce precise numerical targets for CO2 emissions in all regions for the rest of the century, subject to political constraints: subject to political constraints: No country suffers loss (PDV) > Y=1% GDP, by signing up ex ante, nor in any one period suffers a loss > X=5% GDP, by abiding ex post. nor in any one period suffers a loss > X=5% GDP, by abiding ex post.

9. 9 “Is it economics?” Define economics as maximization of objectives subject to constraints. That applies not just to private agents maximizing expected utility subject to budget constraints, but also to how policy-makers can maximize objectives subject to political constraints. Not the same as what other climate modelers do: –cost-benefit analysis (Integrated Assessment models), –or minimizing economic costs subject to the constraint of attaining a given environmental goal.

10. 10 Maximizing the credibility of agreement, for any given environmental goal Aggressiveness of targeted cut in CO2 concentrations by 2100 Credibility of an agreement, Vs. probability that it will un- ravel because (e.g.) some key players find that complying imposes huge economic costs, relative to dropping out. Frankel (2009) Bosetti & Frankel (2010) Some proposals 500 ppm | 450 ppm | 350 ppm |

11. 11 Stage 2: When the time comes for developing country cuts, targets are determined by a formula incorporating 3 elements, designed so each is asked only to take actions analogous to those already taken by others: Stage 2: When the time comes for developing country cuts, targets are determined by a formula incorporating 3 elements, designed so each is asked only to take actions analogous to those already taken by others: –a Progressive Reduction Factor, –a Latecomer Catch-up Factor, and –a Gradual Equalization Factor. Stage 1: Annex I countries commit to the post-2012 targets that their leaders have already announced. Others commit immediately not to exceed BAU. Proposal

12. 12 The three factors in the formulas Progressive Reduction Factor: –For each 1% difference in income/cap => target is γ % greater emissions abatement from BAU. Latecomer Catch-up Factor: –Gradually close the gap between the latecomer’s starting point & its 1990 emission levels at λ per year. (Goal: avoid rewarding latecomers for ramping up emissions). –Baseline probably now moved from 1990 to 2005. Gradual Equalization Factor: –In the long run, rich & poor countries’ targets converge in emissions per capita at δ per year. (Goal: equity)

13. 13 Where do the parameters come from? They would be negotiated. But a good start is to use parameters implicit in targets that have already been agreed. The degree of progressivity in the PRF can be estimated from observed pattern –in allocations among countries already agreed ( γ=.14). We estimated Latecomer Catch-up parameter from the speed with which US targets close the gap: current vs. 1990 emission levels –in Lieberman-Warner (2008) & Waxman-Markey bills (2009) => λ =.3 per 5-yr. period. Initially we set speed of Gradual Equalization δ=.1, per 5-yr. budget period (which comes to dominate per capita targets toward the end of the century).

14. 14 The targeted reductions from BAU agreed to at Kyoto in 1997 were progressive with respect to income. Cuts ↑ Incomes → This is how we set the parameter in the Progressive Reductions Factor γ =.14

15. 15 The resultant paths for emissions targets, permit trading, the price of carbon, GDP costs, & environmental effects are estimated by means of the WITCH model of FEEM, Milan, co-authored & applied by Valentina Bosetti.

16. 16 Global peak date ≈ 2035 ◙ Constraints are satisfied: -- No country in any one period suffers a loss as large as 5% of GDP by participating. -- Present Discounted Value of loss < 1% GDP. ◙ In 2009 version, CO2 concentrations level off at 500 ppm in the latter part of the century.

17. 17 The last completed paper (2010) co-authored with Valentina Bosetti was an attempt to see if we could hit CO2 concentrations = 450 ppm –by assuming more aggressive parameters in the formulas.

18. 18 Latest study, currently underway (2011) updates all the estimates to reflect recent developments in the economy, environment, & negotiations, –particularly the Copenhagen-Cancun country targets, –and to reflect new technologies Including bio-energy (BE) with Carbon Capture & Storage (CCS). and again tries to attain more aggressive targets. “A Politically Feasible Architecture for Global Climate Policy: Specific Formulas and Emission Targets to Build on Copenhagen & Cancun” –with Bosetti –for the UN.

19. 19 2011 EU27 + 20 other countries

20. 20 Progressivity in the Cancun numbers setting “hot air” to 0 for 6 FSU countries a a Regression line γ =.13 t =3.9 R 2 =.44 Emissions targets for 2020 expressed vs. BAU (WITCH model) Cuts ↑ The implicit progressivity coefficient is almost exactly the same as the one we have been using:.13 ≈.14 ! => external validation of the political economy of approach 2011

21. 21 Our 11 regions: EUROPE = –Old Europe + –New Europe US = The United States KOSAU = Korea & S. Africa & Australia (3 coal-users) CAJAZ = Canada, Japan & New Zealand TE = Russia & other Transition Economies MENA = Middle East & North Africa SSA = Sub-Saharan Africa SASIA= India + the rest of South Asia CHINA = PRC EASIA = Smaller countries of East Asia LAC = Latin America & the Caribbean

22. 22 Figure 2: Global emission targets resulting from the formulas & parameters under the 500 ppm goal Using Cancun targets, near term cuts are bigger than in our earlier work.

23. 23 Fig.3: Targets & emissions by OECD countries under the 500 ppm goal Predicted actual emissions exceed caps, by permit purchases. }

24. 24 Fig.4: Targets & emissions, developing countries under the 500 ppm goal Predicted actual emissions fall short of caps, by permit sales. }

25. 25 Figure 8: Effect on energy prices, under 500 ppm goal

26. 26 Figure 5: Global economic costs (% of income) in PDV under 500 ppm goal (without BE-CCS) Global cost: close to 1% of income

27. 27 USA EU KOSAU CAJAZ TE MENA SSA SASIA CHINA EASIA LAM INDIA 1.1% 0.6% 2.5% 1.2% 2.7% 3.3% -0.1% 0.0% 1.9% 1.1% 1.4% -1.7% Regional Cost measured with respect to baseline (no climate policy) Regional Cost measured with respect to case where coalition continues and country free rides. USA EU KOSAU CAJAZ TE MENA SSA SASIA CHINA EASIA LAM INDIA 1.1% 0.5% 2.5% 1.3% 2.9% 4.2% -0.1% -0.3% 1.6% 0.7% 1.4% -1.7% Economic cost to each country/region (Net Present Value of income losses) Cost is particularly high to oil producers – even if they drop out.

28. 28 Figure 7a: Economic losses of each region, (% of income) under 500 ppm goal, 2010-2045

29. 29 Figure 7b: Economic losses of each region, (% of income) under 500 ppm goal, 2050-2090

30. 30 Figure 11: Path of concentrations hits the 500 ppm CO2 goal

31. 31 Figure 12: Rise in Temperature under the 500 ppm CO2 goal 3°C vs. 4 ° C under BAU

32. 32 Summary Our framework allocates emission targets across countries in such a way that every country feels it is only doing its fair share corresponding to what others have done before it, taking due account of differences in income, and minimizing chance that any country will bear a cost above threshold, Otherwise, announcements of distant future goals are not credible, will not have the desired effects. –This framework—in providing for a decade-by-decade sequence of emission targets, each determined on the basis of a few principles and formulas— –is flexible enough to accommodate, by small changes in the formula parameters, changes in circumstances during the century.

33. Papers a vailable at: http://ksghome.harvard.edu/~jfrankel/currentpubsspeeches.htm On Climate Change http://ksghome.harvard.edu/~jfrankel/currentpubsspeeches.htm

34. 34 Appendices 1) Results from the 2009 paper (target = 500 ppm) 2) Results from 2010 paper (target = 460 ppm) 3) Extension for future work: Uncertainty 4) Trade policy: In the meantime, before all countries participate, what border measures are appropriate regarding carbon-intensive imports?

35. 35 Dates at which developing countries were asked to limit emissions to BAU: –Lat Am 2010 –MENA 2010 –China 2010 –India 2010 –SEA sia 2010 –Africa 2025 to cut below BAU, for 500 ppm goal: –2035 –2025 –2030 –2050 to cut emissions below BAU, for 460 ppm goal: –2020 –2025 –2050 (2009)

36. 36 Emissions path for rich countries Fig. 2b Predicted actual emissions exceed caps, by permit purchases. A property I have come to regard as useful: permits are a relatively small share of reductions. } (2009)

37. 37 Emissions path for poor countries Fig. 4b Predicted actual emissions fall below caps, by permit sales. } (2009)

38. 38 Price of Carbon Dioxide Fig. 6b rises slowly over 50 years, then rapidly. (2009)

39. 39 Concentrations stay below 500 ppm goal Fig. 7b (2009)

40. 40 Temperature rises 3° rather than 4° Fig. 8b Yes, I know. The pay-off is a let-down. (2009)

41. 41 Appendix 2: Results from Bosetti & Frankel (2010) Attempt to attain 450 ppm

42. 42 Bottom line The best we could do was attain 460 ppm Even then, we had to loosen our political/economic constraints: –We had to raise the threshold of costs above which a country drops out, as high as Y =3.4% of income in PDV terms, –and X =12 % in the worst budget period. (2010)

43. 43 Target allocations to hit goal of 460 ppm Actual Emission per capita throughout the century, for 11 regions source: Bosetti & Frankel (Feb.2011) source: Bosetti & Frankel (Feb.2011) Figure 2 (2010)

44. 44 Figure 3: Assigned targets & actual emissions for industrialized countries, aggregate 460 ppm (Note: Predicted actual emissions exceed caps by permit purchase amounts.) (2010)

45. 45 Figure 4: Assigned targets & actual emissions for poor countries, aggregate 460 ppm (Note: Predicted actual emissions fall below caps by permit sales amounts) (2010)

46. 46 Figure 5: Assigned targets & actual emissions for all countries, aggregate Goal: 460 ppm concentration of CO2 in year 2100 (2010)

47. 47 Figure 7: Income Losses by Region and Period over the Century (460 ppm) 7 b) 2050- 2100 7 a) 2010- 2045

48. 48 Figure 8: Global Income Loss -- by Budget Period, 2010-2100, and PDV (discounted to 2005) (2010)

49. 49 Figure 9:Losses by Region -- PDV (discounted to 2005 at 5% discount rate), 2010-2100 (2010)

50. 50 Figure 10: CO2 concentrations to achieve year-2100 goal of 460 ppm (2010)

51. 51 Figure 11: Rise in temperature under proposed targets (460ppm) vs. BAU Even though the 460 ppm target is achieved by mid-century, the pay-off in further temperature moderation, relative to 500 ppm, is not large. There are diminishing returns to CO2 abatement in two senses: The marginal cost of abatement rises in dollar terms, and the marginal cost of temperature moderation rises in terms of CO2. (2010)

52. 52 Figure A1. Choosing country targets to minimize threshold for PDV country costs loses the simplicity of a common formula for all, (green triangles) without much gain in reducing PDV of global losses

53. 53 Figure A2. Varying the developing country start dates tightens or loosens the CO2 concentration objective (blue diamonds)

54. 54 Conclusions Some may conclude that the goals of 380 or 450 ppm in CO2 concentrations are not attainable in practice, –and that our earlier proposal for 500 ppm is the better plan (Frankel, 2009). –We take no position on the best environmental goal. –Rather, we submit that, whatever the goal, our formulas will give targets that are more practical economically and politically than approaches that have been proposed by others. (2010)

55. 55 Appendix 3: More on the current results (2011) Details on updating Details on the 500 ppm case A more aggressive climate mitigation goal

56. 56 Revisions in the WITCH model OIL MARKET: is now fully represented. This increases the cost for oil exporters, & so globally. TECHNOLOGIES: –Gas with CCS is now a technological option, –Wind is now modeled given latest numbers, –Bioenergy with CCS can be included. DATA: The model has been recalibrated to reproduce most updated dataset for economy & technology. CLIMATE: The climate module has been updated. We account better for aerosols (though still a big incognita).

57. 57 a a Regression line γ =.16 t =3.7 Cuts ↑ Emissions targets for 2020 expressed vs. BAU (WITCH model) R 2 =.42 Progressivity in the Cancun numbers letting the 6 FSU countries keep their “hot air” 2011

58. 58 How targets are assigned (500 ppm case) 20202050 EU.-30% below 1990 emissionsprogress to a 50% below 1990 USA.-17% below2005.-83% below 2005 Australia, SAf & SK.-34% below baseline.-50% below baseline Japan, NZ & Canada).-25% below 1990as before TE.min(-25% below 1990 or BAU).min(-50% below 1990 or BAU) LAM BAUCap based on parameters India BAU EASIA BAUCap based on parameters SASIA BAU, with REDDBAU CHINA.-8.5% wrt BAUCap based on parameters SSA BAU, with REDDBAU MENA BAUCap based on parameters 2011

59. 59 Figure 6: Global economic costs (% of income) under 500 ppm goal (with BE & CCS) 2011

60. 60 Figure 10: Global emissions if only one country drops out, but cooperation otherwise continues 2011

61. 61 How targets are assigned in the aggressive case (to attain CO2 concentrations of 450 ppm) 20202050 EU.-30% below 1990 emissionsprogress to 50% below 1990 USA.-17% below2005.-83% below 2005 Australia, SA and SK.-34% below baseline.-50% below baseline Japan, NZ & Canada.-25% below 1990as before TE.min(-25% below 1990 or BAU).min(-50% below 1990 or BAU) LAMBAUCap based on parameters IndiaBAUCap based on parameters EASIABAUCap based on parameters SASIABAU, with REDDCap based on parameters CHINA.-8.5% wrt BAUCap based on parameters SSABAU, with REDDBAU MENABAUCap based on parameters 2011

62. 62 Figure 15: Global emission targets under the more stringent goal 2011

63. 63 Figure 13: Targets & emissions among OECD countries under the more stringent goal 2011

64. 64 Figure 14: Targets & emissions among developing countries under the more stringent goal 2011

65. 65 Figure 16: Emissions in per capita terms, country by country, period by period 2011

66. 66 Figure 19: Energy costs in stringent case 2011

67. 67 Figure 17: Global economic costs (% of income) under stringent goal with BE & CCS 2011

68. 68 Fig.18a: Economic losses of each region (% of income) under stringent goal, 2010-2045 2011

69. 69 Figure 18b: Economic losses of each region, (% of income) under stringent goal, 2050-2090 2011

70. 70 Figure 21a: Concentrations of CO2 under the more stringent goal (including Bioenergy with Carbon Capture & Storage) 2011

71. 71 Fig.21b: Concentrations of all 6 greenhouse gases under the more stringent goal (including Bioenergy and Carbon Capture & Storage)

72. 72 Figure 22: Rise in Temperature under stringent concentrations goal, with BE & CCS 2011

73. 73 Fig. 20: Size of the international emission permit market

74. 74 Fig.20b: Permit trade never exceeds 1/3 of baseline for any country, a desirable property

75. 75 Appendix 4: Future research Most important extension: introduce uncertainty, re: –Economic growth rates –Technological progress (carbon intensity at a given price) –Weight placed by political process on enviro objective Robustness in design of framework regarding losses: –Periodic renegotiation of parameters –& within-decade indexation of numerical targets to GDP. Show that the approach is desirable even if the model’s cost estimates turn out to be too pessimistic –(or too optimistic), –because early steps will be very similar regardless.

76. 76 Appendix 5: Trade Will the global climate regime collide with the global trade regime?

77. 77 Leakage and lost competitiveness = w hat “unfairness” means in rich countries Could border measures (tariffs, permit requirements for imports…) give non-participants an incentive to join, and level the playing field in the meantime? –Yes, in theory. –If properly designed, they could even be WTO-consistent, building on precedents: Montreal Protocol, shrimp-turtle decision, Brazil tires, and D-G Lamy speeches (2 Nov., 2009). –In practice, however, border measures written by politicians will almost certainly be discriminatory, WTO-inconsistent, and undermining of climate goals.

78. 78 What form should border measures take? 1.Best choice: multilateral sanctions. (Not likely.) 2.Next-best: national sanctions adopted under multilateral guidelines 1.Penalties can only be applied by participants-in-good standing 2.Judgments to be made by technical experts, not politicians 3.Penalties only in 6 or 8 of the most relevant sectors (aluminum…). 3.Third: No border measures. 4.Each country chooses carbon tariffs as it sees fit. 5.Worst choice: national measures are subsidies ( e.g., free permits) to adversely affected firms.