#1California and Guangdong: A tale of two cap-and-trade programs Pu Wang Cheng-Kuan Lin Yi-Hua Wu#2Guangdong: Largest province in population and GDP • Pop (2015): 108.49 million GDP (2015): $1.07 trillion Area: 69,410 sq mi • Most progressive province in China • Hub of high tech industries in China: Tencent (WeChat), Huawei, Foxconn... Leader in low-carbon policy: Guangzhou and Shenzhen to peak CO2 emission by 2020 and 2022. • Launched pilot emission trading system(ETS) in 2013 California: • • • Largest state in population and GDP • Pop (2015): 39,14 million GDP (2015): $2.46 trillion Area: 163,696 sq mi Most progressive state in the U.S. Hub of IT industries in the U.S.: Google, Apple, Tesla... Leader in low-carbon policy: emission reduction goals in 2020 and 2050 Launched cap-and-trade program in 2013#3• Research ideas: A comparison between the design and implementation of Guangdong and California's cap-and-trade programs; Using model simulations to examine how emissions in different sectors were affected by the C/Ts; • How the different socioeconomic settings and policy designs influence their effectiveness; . California's program is considered as the best designed of its kind; find the gap between Guangdong and CA could help. improve C/T in China.#4Differences in economic structure CA Wholesale, Retail 372Mton CO2 (2014) Trade and Agriculture Catering Services 1% Transportation and communication 44% 1% Petroleum Processing Buildings 7% Electricity and heat imports 10% Petroleum Wholesale, Retail Trade and Catering Services 3% Transportation and communication Electricity and heat in State 17% Processing Primary metals 1% 8% Glass and cement 1% 8% Primary metals Oil and Gas Extraction 0% 5% 14% GD 444Mton CO2 (2015) Agriculture 1% Buildings 8% Glass and cement 11% Electricity and heat in State 50%#5Legislation California ⚫ In June 2005, Governor Schwarzenegger signed Executive Order S305, requiring the state to reduce its greenhouse gas (GHG) emissions levels to 2000 levels by 2010, to 1990 levels by 2020, and to a level 80% below 1990 levels by 2050. • AB 32, or the California Global Warming Solutions Act of 2006, gave authority to California Air Resources Board (ARB) to develop regulations and market mechanisms to reduce GHG to 1990 levels by 2020. Guangdong . ⚫ The National Development and Reform Commission (NDRC) published "The notice of developing carbon emission trading pilot programs" in 2011. . ETS in Guangdong is one of the seven pilot programs in order to provide lessons to future national program. Subsequently, Guangdong DRC published "Carbon emissions management interim measures for Guangdong Province". • Lack of detailed scrutiny, policy certainty, and support for enforcement.#6Seven ETS pilot programs in China Chongqing Hubei Beijing -Tianjin Shanghai Guangdong Shenzhen#7CA Complementary policies Recommended Reduction Measures Reductions Counted Towards 2020 Target (MMTCO₂E) ESTIMATED REDUCTIONS RESULTING FROM THE COMBINATION OF CAP- AND-TRADE PROGRAM AND COMPLEMENTARY MEASURES California Light-Duty Vehicle Greenhouse Gas Standards Implement Pavley standards Develop Pavley II light-duty vehicle standards Energy Efficiency Building/appliance efficiency, new programs, etc. Increase CHP generation by 30,000 GWh Solar Water Heating (AB 1470 goal) Renewables Portfolio Standard (33% by 2020) Low Carbon Fuel Standard Regional Transportation-Related GHG Targets Vehicle Efficiency Measures 31.7 26.3 21.3 15 5 4.5 3.7 146.7 Goods Movement Ship Electrification at Ports System-Wide Efficiency Improvements Million Solar Roofs 2.1 Medium/Heavy Duty Vehicles Heavy-Duty Vehicle Greenhouse Gas Emission Reduction 1.4 (Aerodynamic Efficiency) Medium- and Heavy-Duty Vehicle Hybridization High Speed Rail 1.0 Industrial Measures (for sources covered under cap-and-trade program) Refinery Measures 0.3 Energy Efficiency & Co-Benefits Audits Additional Reductions Necessary to Achieve the Cap 34.4#8GD Complementary policies. From 2011 to 2015, China implemented the "ten thousand firms energy saving and low-carbon action" program. The overall goal was to reduce energy consumption per unit of GDP by 16%, and reduce CO2 emission per unit of GDP by 17%. • A command-and-control approach: the participating firms were required to improve their energy efficiency through upgrading technologies, optimizing production processes, and shutting down inefficient facilities.#9Cap setting California • Returning to 1990 level by 2020 (431 million ton) • The 2014 level was 441 million ● ton. Cap decreases 2% annually from 2012-14, 3% annually thereafter. Guangdong • No explicit cap or target. • A bottom-up approach: cap is calculated from the sum of facility-level data. . • No specific future trajectories; likely to adjust the cap according to economic conditions. Energy intensity target: reduce energy consumption per GDP unit by 18% between 2010-15. • Carbon intensity target: reduce CO2 emission per GDP unit by 19.5% between 2010-15.#10CA emission reduction targets MMTCO₂E 700 600 500 Total Emissions: 596 MMTCO2E Agriculture High GWP Recycling & waste Industry 400 Natural gas 300 Electricity 200 Reduction Measures Reductions from uncapped sectors: Total reductions of 27.3 MMT Industrial measures: 1.1 MMT High GWP measures: 20.2 MMT Recycling & waste: 1.0 MMT Sustainable forests: 5.0 MMT Reductions from capped sectors: Total reductions of 146.7 MMT (including 112.3 MMT from specified measures): Pavley standards: 31.7 MMT Energy efficiency: 26.3 MMT 33% RPS: 21.3 MMT Total Emissions 422 MMTCO2E Cap is set at 365 MMT Agriculture High GWP Recycling & waste LCFS: 15.0 MMT Regional targets: 5.0 MMT Vehicle efficiency: 4.5 MMT Capped sectors 3.7 MMT 2.1 MMT Goods movement: Million solar roofs: Transportation 100 Heavy/medium veh: 1.4 MMT Industrial measures: 0.3 MMT High speed rail: 1.0 MMT 0 Business-as-Usual Scoping Plan#11Coverage California • Covers 7 types of GHG Guangdong • Only CO2 is covered • Program covers about 450 entities . • Starts in 2013 for electricity generators and large industrial facilities emitting 25,000 MTCO2e or more annually Starts in 2015 for distributors of transportation, natural gas, and other fuels Imported electricity is . covered. • · Program covers 193 entities (2014) Covered 4 industries: electricity generation, petrochemical(refinery), cement, Steel; The threshold is 20,000 ton annually. Transportation not covered. . Imported electricity not covered.#12Allowance allocation CA • Free allocation early in the program, transitions to more auction later in program Allocation of allowances for most industrial sectors is set at about 90 percent of average emissions, based on benchmarks that reward efficient facilities • For most industrial sectors, distribution of allowances is updated annually according to the production at each facility Offset credits: Allowed for up to 8 percent of a facility's compliance obligation . GD A combination of benchmarking (primary), grandfathering, and auctioning • Auction: 5% in power sector, 3% in other sectors; the share of auctioning should increase overtime. . Allowances are updated annually. • Offset credits: Allowed for up to 10 percent of a facility's compliance obligation#13Data GDP California Guangdong and Jiangsu 2000-2014, 9 sectors 2000-2015, 9 sectors Energy consumption 2000-2014, 29 sectors 2000-2015, 29 sectors CO2 emissions 2000-2014, 29 sectors, published 2000-2015, 29 sectors, calculated by ARB from energy consumption data Facility level data Emissions from each covered facility Guangdong: List of the names of the firms covered by cap-and- trade Energy balance sheets (imports and exports) Imports data available 2000-2014#14Trends in CO2 emissions. 800 700 600 500 400 300 200 100 MT CO2 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 CA GD JS#15Generalized mixed-effects model • estimate the carbon emission and carbon intensity before and after implementing cap-and trade, stratified by different industrial sectors. • We performed the PROC MIXED procedure using SAS version 9.4 (SAS Institute, Cary, NC, US) to estimate the effect of selected factors on carbon dioxide emission. • * E[Yij Xij] = Po + B₁ * Cap&Trade₁j + ß2 * state; + ẞ3 * year; + ẞ4 * state¡ * ß3 ß year; ++ẞ5 * I(Sector); + ß6 * GDP; + ẞ7 * state; * I(Sector)i Cap&Tradejj ⚫i: industry sectors; j: time in years since the baseline year (from 1 to 6); . . . E[Y_ij |X_ij]: the expected carbon dioxide emission conditioned on covariates X; Bo: the intercept for the effect; B1 to B6 are estimated coefficients for the marginal effects on the emission derived from maximum likelihood method. Cap&Trade are dummy variable, state are categorical variable for the three states/provinces, and I(Sector) is indicator variable for the six sectors.#16• . Similar model with outcome as carbon intensity (KgCO2/GDP) is expressed as following: E[λij|Xij] = ßo + ß₁ * Cap&Trade¡¡ + ß2 * state; + ß³ * year¿ + ß * state; * year; + +ẞ5 * I(Sector); + ß6 * state; * I(Sector); * Cap&Trade{j • Where E[λ_ij |X_ij] denotes the expected carbon intensity conditioned on covariates X.#17Solution for Fixed Effects Effect state ct sectorn Estimate Standard t Value Pr>|t| Error Intercept -48.7684 12.5163 -3.90 0.0002 ct ct 1 0 -6.6692 13.5532 -0.49 0.6241 0 GDP_mil 0.000471 0.000107 4.43 <.0001 yrp 0.08729 1.7698 0.05 0.9608 yrp*state CA -5.4866 2.6336 -2.08 0.0407 yrp*state GD -4.9747 2.6330 -1.89 0.0628 yrp*state JS 0 state* ct sectorn CA 1 1 152.75 45.4345 3.36 0.0012 state* ct sectorn CA 1 2 -292.31 45.1333 -6.48 <.0001 state* ct sectorn CA 1 3 137.67 41.8532 3.29 0.0015 state* ct sectorn CA 1 4 181.20 27.3039 6.64 <.0001 state*ct sectorn CA 1 5 -100.71 27.0329 -3.73 0.0004 state* ct sectorn CA 1 9 1.1793 18.9168 0.06 0.9505 state* ct sectorn CA 0 1 132.65 37.7246 3.52 0.0008 state* ct sectorn CA 0 2 -294.32 40.3577 -7.29 <.0001 state* ct sectorn CA 0 3 120.07 34.3428 3.50 0.0008 state* ct sectorn CA 0 4 177.91 20.2524 8.78 <.0001 state* ct sectorn CA 0 5 -85.3285 19.1354 -4.46 <.0001 state* ct sectorn CA 0 9 0 state* ct sectorn GD 1 1 2.8302 16.2562 0.17 0.8623 state* ct sectorn GD 1 2 -284.71 15.4725 -18.40 <.0001 state* ct sectorn GD 1 3 12.1169 16.7617 0.72 0.4720 state* ct sectorn GD 1 4 28.1017 15.9431 1.76 0.0821 state* ct sectorn GD 1 5 0.1707 15.3621 0.01 0.9912 state* ct sectorn GD 1 9 0 state* ct sectorn GD 0 1 -14.6268 12.0679 -1.21 0.2294 state* ct*sectorn GD 0 2 -235.68 12.0611 -19.54 <.0001 state* ct*sectorn GD 0 3 -5.5244 12.2152 -0.45 0.6524 state* ct sectorn GD 0 state* ct*sectorn GD 0 state* ct sectorn GD 0 459 5.8596 12.0332 0.49 0.6277 -10.3111 12.0573 -0.86 0.3952 0#18Solution for Fixed Effects Effect state ct sectorn Estimate Standard t Value Pr>|t| Error Intercept ct ct 1 0 0.1725 0.07518 0.02520 6.84 <.0001 0.04788 1.57 0.1206 0 yrp -0.01316 0.004620 -2.85 0.0057 yrp*state CA 0.003477 0.009449 0.37 0.7139 yrp*state GD -0.02786 0.009449 -2.95 0.0043 yrp*state JS 0 state* ct sectorn CA 1 state* ct*sectorn CA 1 state* ct sectorn CA 1 state* ct sectorn CA 1 state* ct sectorn CA 1 state* ct sectorn CA 1 state* ct sectorn state*ct sectorn state* ct sectorn state* ct sectorn CA 0 CA 0 CA 0 123459123 -0.04683 0.07302 -0.64 0.5232 -1.2686 0.07302 -17.37 <.0001 -0.01771 0.07302 -0.24 0.8090 -0.3167 0.07302 -4.34 <.0001 -0.01113 0.07302 -0.15 0.8793 -0.05934 0.06771 -0.88 0.3836 -0.00478 0.04322 -0.11 0.9122 -1.1539 0.04322 -26.70 <.0001 0.02318 0.04322 0.54 0.5934 CA 0 4 -0.1323 0.04322 -3.06 0.0031 state* ct sectorn CA 0 5 0.03568 0.04322 0.83 0.4117 state*ct sectorn CA 0 9 0 state* ct*sectorn GD 1 1 -0.1178 0.05467 -2.16 0.0344 state* ct sectorn GD 1 2 -0.9720 0.05467 -17.78 <.0001 state* ct*sectorn GD 1 3 -0.03150 0.05467 -0.58 0.5662 state* ct sectorn GD 1 4 0.4100 0.05467 7.50 <.0001 state* ct sectorn GD 1 state* ct sectorn GD 1 state*ct*sectorn state* ct sectorn GD 0 GD 0 state* ct sectorn state* ct sectorn state* ct sectorn state* ct sectorn GD GD 0 GD 0 GD 0 59123459 0.01420 0.05467 0.26 0.7957 0 -0.1525 0.04322 -3.53 0.0007 -0.7454 0.04322 -17.25 <.0001 -0.06158 0.04322 -1.42 0.1583 0.6494 0.04322 15.02 <.0001 -0.02937 0.04322 -0.68 0.4988 0#19Industry MtCO2 600- MtCO2 2.5 692 634 640 × 596 579 2.0 530 Emission reduction after 2 years of cap-and-trade × 1.5 400- 417 Emissions ✗ 398 377 390 395 Actual BAU Carbon intensity Actual BAU 324 CA -3.30% 0.47% -11.93% -2.65% 1.0 GD -17.05% 6.61% -26.27% -3.40% 200- 104 107 103 170 166165 JS 9.22% 0.49% -0.52% -0.17% 0.5 0- state -4 Observed year 0.0 5 6 CA X GD × JS#20MtCO2 Transportation MtCO2 2.5 X 165 × 161 150- 158 158 157 159 100- 50- 53 53 0- 2.0 Emission reduction after 2 years of cap-and-trade 1.5 Actual BAU 1.0 CA Emissions Carbon intensity Actual BAU 0.17% -0.83% -12.60% -11.94% 63 57 54 55 57 GD 14.42% 5.68% -4.27% -6.60% 27 30 8x 34 34 38 39 JS 14.17% 4.82% 2.35% -2.05% 0.5 ½ 4 -5 Observed year state × CA × GD X JS 16 0.0#21Wholesale, Retail CA 372Mton CO2 (2014) Trade and Agriculture Catering Services 1% Transportation and communication 44% 1% Petroleum Processing Buildings 7% Electricity and heat imports 10% Petroleum Wholesale, Retail Trade and Catering Services 3% Transportation and communication GD 444Mton CO2 (2015) Agriculture 1% Buildings 8% Electricity and heat in State 17% Processing Primary metals 1% 8% Glass and cement 1% 8% Primary metals Oil and Gas Extraction 0% 5% 14% Glass and cement 11% Electricity and heat in State 50%#22Other sectors MtCO2 40- MtCO2 38.49 39.01 38.77 37.1 36.12 36.82 X 34.37 33x71 31.97 33.00 31.98 30- 27.14 00 0.4 Emission reduction after 2 years of cap-and-trade 0.3 Emissions 0.2 Carbon intensity 20- Actual BAU Actual BAU 18.34 16.8 16.48 17.28 CA -11.46% -15.79% 14.8 13.68 0.1 GD 21.27% -3.00% 10- JS 11.29% -18.75% 0- i 2 state 3 4 Observed year × CA × GD × JS -CO 6 -5 0.0#23Principal component analysis We adopt principal component analysis to forecast CO2 emissions for 29 sectors in CA, GD, and JS, respectively. • We divide data sets into two categories: the first contains economic variables, and the second includes CO2 emissions of detail industrial sectors. The economic variables include value added output of each industrial sector in the i region - where i denotes California, JS, and GD. The CO2 emissions variables include emissions from each of the 29 sectors.#24. • Let X (i,t)^ECON denote a vector of value added output in region i at period t, and Y_(i,t)^CO2 denote a vector of CO2 emissions in region i at period t: XECON = [Xi,1t, Xi,2t, ..., Xi,nt] i,t • And YC02 . = i,t [Yi,1t, Yi,2t, Yi,mt] ***) where n is the number of value added output in region i, and m is the number of emission sources, and t is the sample period (t = 1, ..., T).#25. Use principal component analysis to estimate factors from economic variables and CO2 emissions, respectively. We follow the principal component analysis of Stock and Watson (2002). Let FECON denote the factor for XEE and FC02 denote the factor for YC02. The factor model is ECON , i,t • Xi,jt = FEcon AĘcon x + ejt • Yi,jt FC02 2002 + et = CO2 where j = 1,...,n, and i = 1, ..., m. e) and et and et are idiosyncratic errors for xi,jt and Yi,jt, respectively. The objective function for the estimation of FEcon and 2 Econ is min n FECO con NT Σj=1&l=1(xi‚jt — FЕcon 2Econ)² FEcon, and that for FC02 and 1902 is · min CO2 CO2 NT Σ1Σ-1(Vijt - FC021f02)2 The estimation strategy for Econ and ^ Econ as well as CO2 and 102 can be found in Bai and Ng (2002).#262. Forecast the future trend of factors using the vector autoregressive (VAR) model. • Write the VAR model as: • = Ft CF-1B + Ut t • where Ft = [FC02, FECON] - the vector of estimated factors, c is a constant term, and B is a parameter matrix. The lag order of VAR is set to be one, due to the short sample period. We can forecast the future trend of factors using the equation above for periods t = T + 1, ..., T + h.#273. Use the future trend of factors to forecast the CO2 emissions of detail industrial sector . • We estimate the relationship between emission of each sector and factors using: . = Yi,jtc+FA + ut • where A is a parameter matrix and u₁ is measurement error. Once we have estimated ĉ and Â, we can estimate the future movements of Yi,jt for period t = T + 1, ..., T+h, given the forecast of Ft. ==#2831 30 29 28 27 26 25 24 2000 2002 2004 2006 2008 2010 2012 2014 8 7 6 5 4 3 CA_Petroleum Processing Mean CI CI 2 1 0 2000 2002 2004 2006 2008 2010 2012 2014 GD_Petroleum Processing Mean Cl CI Two years after Cap-and-trade compared to BAU: CA: -3.31% GD: -13.5%#296 5 4 3 2 1 0 2000 2002 2004 2006 2008 2010 2012 2014 70 CA_Glass and cement Mean CI 60 50 40 30 20 10 0 2000 2002 2004 2006 2008 2010 2012 2014 GD_Glass and cement Mean Cl CI Two years after Cap-and-trade compared to BAU: CA: 14.6% GD: -10.28%#301.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 2000 2002 2004 2006 2008 2010 2012 2014 60 60 510 50 40 30 20 10 CA_Primary metals Mean CI 0 2000 2002 2004 2006 2008 2010 2012 2014 GD_Primary metals Mean CI CI Two years after Cap-and-trade compared to BAU: CA: 12.11% GD: -21.84%#3180 70 60 50 40 30 20 10 0 T 2000 2002 2004 2006 2008 2010 2012 2014 CA_Electricity and heat in State Mean 350 300 250 200 150 100 50 0 2000 2002 2004 2006 2008 2010 2012 2014 GD_Electricity and heat Mean CI CI CI Two years after Cap-and-trade compared to BAU: CA: -2.2% GD: -21.51%#3220 2652322° 70 60 50 40 30 10 0 2000 2002 2004 2006 2008 2010 2012 2014 CA_Electricity and heat imports Mean 2013 2014 2015 GD import 1175 1709 1426 CI Two years after Cap-and-trade compared to BAU: CA: -4.32% GD: N/A#33185 180 175 170 165 160 155 150 145 140 2000 2002 2004 2006 2008 2010 2012 2014 CA_Transportation and communication Mean CI CI 40 70 60 50 265222 30 20 10 0 2000 2002 2004 2006 2008 2010 2012 2014 GD_Transportation and communication Mean CI Two years after Cap-and-trade compared to BAU: CA: 1.16% GD: 11.16%#3434 29 25 322202 33 31 28 27 26 2000 2002 2004 2006 2008 2010 2012 2014 25125250 40 35 30 CA_Buildings Mean CI 4 2000 2002 2004 2006 2008 2010 2012 2014 GD_Buildings Mean CI CI Two years after Cap-and-trade compared to BAU: CA: -18.58% GD: 16.05%#35Summary of PCA results . · All GD's covered heavy industrial sectors had dramatic decrease in emissions: 10-20%. This would make the cap unbinding, so it was more likely to be caused by economic slowdown. CA's heavy industries were relatively small, and some of them had noticeable increase in emissions. But overall the C/T achieved its reductions, mostly from imported electricity. Emissions from transportation were not affected in either CA or GD-covering refinery is not sufficient; • GD imported electricity increased, while CA's decreased: the importance to address leakage in electricity grids. • Emissions from buildings: CA decreased, while GD increased: indicating GD's trend to transition from manufacturing economy to service economy.#36Conclusions • Two years after implementation of C/T, both CA and GD experienced significant emission reduction, but the patterns are different; • Emissions in all heavy industries in GD dropped dramatically, but not true for heavy industries in CA; • CA emission reduction mostly comes from electricity, particularly imported electricity CA and GD face different types of challenges in emission reduction; • CA is a relatively stable economy → smooth emissions trend, easier for cap setting. • GD is a transitioning economy; its heavy industries are susceptible to economic fluctuations → cap setting is difficult; • GD: emissions in heavy industries are declining overall, but emissions from transportation, residential, and commercial activities are increasing: C/T is not sufficient to address these sources.#37Conclusions (2) The importance to tailor C/T system for the specific socioeconomic conditions • GD's current system almost exactly copied the designing rules of EU and CA Other pilot programs, most importantly, the national C/T program, are likely to copy the western-style program design Lessons China should learn: emission dataset, transparent rules, legislation support, enforcement... Changes China needs to make: coverage, allowance allocation (equity. concerns, declining industries), trading rules (State owned companies), complementary polices (transportation, buildings, electricity system) • It is still too soon to evaluate the effectiveness of the two programs . • CA: transportation started to be covered in 2015 • GD: economic recovery in 2016 would change the emissions trend; . a longer time series data would allow separation of the effects of economic change and C/T.