Laurent Drouet and Johannes Emmerling 1. IAMC Meeting, Nov 16th, 2015, Potsdam. Climate policy under socio-economic scenario uncertainty

Transcription

1 Climate policy under socio-economic scenario uncertainty Climate policy under socio-economic scenario uncertainty Laurent Drouet and Johannes Emmerling 1 1 Fondazione Eni Enrico Mattei (FEEM) and CMCC IAMC Meeting, Nov 16th, 2015, Potsdam Johannes Emmerling FEEM IAMC 2015, Potsdam 1 / 16

2 Climate policy under socio-economic scenario uncertainty Introduction Motivation Socioeconomic Drivers of the ve SSPs Climate policies under very dierent baseline assumptions Policy costs or impacts under dierent GDP and population scenarios Johannes Emmerling FEEM IAMC 2015, Potsdam 2 / 16

3 Climate policy under socio-economic scenario uncertainty Introduction Motivation Socioeconomic Drivers of the ve SSPs Climate policies under very dierent baseline assumptions Policy costs or impacts under dierent GDP and population scenarios Research questions 1 What is a robust and optimal climate policy under baseline uncertainty? 2 How can policy costs and other measures be compared if dierent baselines are considered? 3 How can scenario development exercises inform decisions under uncertainty taking into account the dierent worlds described Johannes Emmerling FEEM IAMC 2015, Potsdam 2 / 16

4 Climate policy under socio-economic scenario uncertainty Scenario or baseline uncertainty The Shared Socioeconomic Pathways (SSPs) Consider only dierent baseline assumptions about GDP and population (OECD/IIASA)! Johannes Emmerling FEEM IAMC 2015, Potsdam 3 / 16

5 Climate policy under socio-economic scenario uncertainty Scenario or baseline uncertainty GDP and Population of the SSPs Generate 50 baselines generated from the four extreme SSP baseline as a convex combination of population and GDP per capita 100 SSP3 SSP5 GDP per capita [thousands $2005] SSP1 SSP4 World population [Billions] 10 5 SSP4 SSP5 SSP1 SSP Time Time Johannes Emmerling FEEM IAMC 2015, Potsdam 4 / 16

6 Climate policy under socio-economic scenario uncertainty Scenario or baseline uncertainty GDP and Population of the SSPs Samples pathways and comparison to the SSP2 Middle of the road 100 GDP per capita [thousands $2005] MEAN SSP2 World population [Billions] 10 5 MEAN SSP Time Time Johannes Emmerling FEEM IAMC 2015, Potsdam 5 / 16

7 Climate policy under socio-economic scenario uncertainty Scenario or baseline uncertainty Spatial Heterogeneity of Climate Change Additionally, running 100 carbon budgets ( GtCO 2 ) for each sampled pathway (5,000 runs in total) running the IAM WITCH for each scenario and policy (CB ) (c t,r ) max 1 η w t,r l t,r t,r 1 η 1 (1 + δ) s.t. t t,r emi(c t,r ) CB Johannes Emmerling FEEM IAMC 2015, Potsdam 6 / 16

8 Climate policy under socio-economic scenario uncertainty Scenario or baseline uncertainty Spatial Heterogeneity of Climate Change Additionally, running 100 carbon budgets ( GtCO 2 ) for each sampled pathway (5,000 runs in total) running the IAM WITCH for each scenario and policy (CB ) (c t,r ) max 1 η w t,r l t,r t,r 1 η 1 (1 + δ) s.t. t t,r emi(c t,r ) CB Including damages from climate change (CLIMCOST damage function) 9 India South Asia Sub saharian Africa EU new Consumption loss [%] East Asia Latin America Middle East KOSAU India China USA CAJAZ TE EU Temperature increase since preindustrial [ C] = W (CB;Y,L) Johannes Emmerling FEEM IAMC 2015, Potsdam 6 / 16

9 Climate policy under socio-economic scenario uncertainty Scenario or baseline uncertainty Example of the optimal climate policy Value function for the baseline (α 1,α 3,α 4,α 5 ) = (0,0,1/9,8/9) 1.50 Welfare Carbon budget [GtCO2] Johannes Emmerling FEEM IAMC 2015, Potsdam 7 / 16

10 Climate policy under socio-economic scenario uncertainty Decision framework Decision framework Deterministic case: {Y,L} : CB DET Under uncertainty: expected utility (Y,L) = argmaxw (Y,L,CB) (1) CB CBEU = argmax E 0 W (Ỹ, L,CB) (2) CB Disentangled Epstein-Zin SWF (relative risk acersion ρ η) ( CBDIS = argmax 1 { E 0 (1 η)w (CB;Ỹ CB 1 ρ, L) 1 ρ ) 1 η + 1} 1 (3) Johannes Emmerling FEEM IAMC 2015, Potsdam 8 / 16

11 Climate policy under socio-economic scenario uncertainty Decision framework Decision framework 2.9 density Temperature increase in 2100 [C] Optimal carbon budget [GtCO2] Optimal deterministic CB Risk aversion The eect of risk aversion Johannes Emmerling FEEM IAMC 2015, Potsdam 9 / 16

12 Climate policy under socio-economic scenario uncertainty Decision framework Decision framework - results Optimal carbon budgets for all decision rules Welfare Decision rule DET E0W_CB_DET E0W_CB_DIS E0W_CB_EU Risk aversion Optimal carbon budget [GtCO2] Johannes Emmerling FEEM IAMC 2015, Potsdam 10 / 16

13 Climate policy under socio-economic scenario uncertainty Policy cost measures Policy cost measures (discounted) Policy/damage cost measure: PC = E T t=0 l t(ct pol T t=0 l tc bau t c bau t ) 1 (1+r) t 1 (1+r) t Baseline GDP and population important, choice of discount rate matters. Johannes Emmerling FEEM IAMC 2015, Potsdam 11 / 16

14 Climate policy under socio-economic scenario uncertainty Policy cost measures Policy cost measures (discounted) Policy/damage cost measure: PC = E T t=0 l t(ct pol T t=0 l tc bau t c bau t ) 1 (1+r) t 1 (1+r) t Baseline GDP and population important, choice of discount rate matters. One way to avoid the dependence of dierent future growth scenarios is the use of a Balanced Growth Equivalent (Mirrlees, 1972): Properties: CEBGE = EW (1 η) T t=0 El t (1+g)(1 η)t (1+δ) t dt 1 1 η = CEBGE independent of the assumed growth rate g welfare-based compatible with the IAM framework Collapses to PC for η = 0 and δ = r CEBGE pol CEBGE bau 1 Johannes Emmerling FEEM IAMC 2015, Potsdam 11 / 16

15 Climate policy under socio-economic scenario uncertainty Policy cost measures Policy cost measures across dierent baselines Example: Compare the dierent policy/damage cost measures for impacts from climate change and a two degree target (with impacts!) across dierent baselines summarized by total global GDP Johannes Emmerling FEEM IAMC 2015, Potsdam 12 / 16

16 Climate policy under socio-economic scenario uncertainty Policy cost measures Policy cost measures across dierent baselines Example: Compare the dierent policy/damage cost measures for impacts from climate change and a two degree target (with impacts!) across dierent baselines summarized by total global GDP Damage cost [%] Net policy cost [%] Cost measure CEBGE PC, DR=1% PC, DR=3% PC, DR=5% 6 Cost measure CEBGE PC, DR=1% PC, DR=3% PC, DR=5% Baseline total world consumption over the period [T$] Baseline total world consumption over the period [T$] (a) Damage costs of climate change (b) Policy costs of a two-degree target According to the CEBGE, higher baselines lead to lower damage costs, but higher mitigation costs Johannes Emmerling FEEM IAMC 2015, Potsdam 12 / 16

17 Climate policy under socio-economic scenario uncertainty Optimal climate policy Optimal climate policy and cost measures Compare the dierent measures of net policy costs (mitigation costs and impacts) across the range of policies/carbon budgets considered: 4 2 Net policy cost [%] Cost measure CEBGE PC, DR=1% PC, DR=3% PC, DR=5% Carbon Budget [GtCO 2 ] Johannes Emmerling FEEM IAMC 2015, Potsdam 13 / 16

18 Climate policy under socio-economic scenario uncertainty Optimal climate policy Results Comparison of dierent expected policy cost measures: Decision criterion CB CEBGE PC r=0% PC r=1% PC r=3% E 0 W (CBDET (Ỹ, L);Ỹ, L) varying -0.58% -6.7% -5.2% -2.7% E 0 W (CBEU ;Ỹ, L) % -6.7% -5.2% -2.8% E 0 W (CBDIS ;Ỹ, L), ρ = % -6.9% -5.7% -3.2% Overall, total socioeconomic uncertainty increases the welfare costs from climate change by about 29% (based on the EVPI) Rather balanced evaluation based on the CEBGE, sensitive to discount rate for the PC measures. Johannes Emmerling FEEM IAMC 2015, Potsdam 14 / 16

19 Climate policy under socio-economic scenario uncertainty Conclusion Conclusion Socioeconomic baseline uncertainty plays an important role for the optimal climate policy Comparing policy/damage costs for dierent baseline is not trivial Impact costs seem higher in low baselines, policy costs of stringent policies for high baselines (like SSP5) The SSP provide a useful and broad framework for describing the substantial socioeconomic uncertainties Numerically, we nd that baseline uncertainty increases the welfare cost from climate change by about one fourth Johannes Emmerling FEEM IAMC 2015, Potsdam 15 / 16

20 Climate policy under socio-economic scenario uncertainty Conclusion Conclusion Thank you! Johannes Emmerling FEEM IAMC 2015, Potsdam 16 / 16