Equilibrium Assumptions and Efficiency Agreements: Macroeconomic Effects of Climate Policies for UK Road Transport, 2000

1. Equilibrium Assumptions and Efficiency Agreements: Macroeconomic Effects of Climate Policies for UK Road Transport, 2000-2010 Jonathan Rubin, University of Maine Terry Barker, University of Cambridge

2. Outline Results Intro: macroeconomic models of energy efficiency improvements CGE v. Sectoral models Integration of top-down, bottom-up modelling Modelling the macroeconomic rebound effect using MDM-E3 Application to transport sector Findings and conclusions

3. Results Voluntary efficiency agreement Positive macroeconomic impacts GDP, employment, inflation Reductions in energy demand & CO2 Equivalent fuel taxes (in terms of CO2) Without revenue recycling Lower GDP & employment With revenue recycling that reduces income taxes Eliminates negative macroeconomic impacts Equity/distributional issues remain Inflation neutral scenario, VA & fuel duties

4. UK Energy Efficiency Policies Significant part of 2000 UK Climate Change Programme and 2003 Energy White Paper review of UK CCP launched in Sept 04 - reported March 2006 Updated DTI projections for CO2 emissions (including CCL, 9% renewables, excluding EU ETS): 1990 161 MtC Projection for 2010 144 - 145 MtC (10% reduction) Target for 2010 129 MtC (20% reduction) �Carbon gap� 15 � 16 MtC Department for Trade and Industry has been renamed/converted to Department for Business, Enterprise and Regulatory ReformDepartment for Trade and Industry has been renamed/converted to Department for Business, Enterprise and Regulatory Reform

5. Effect of Climate Change Programme Measures Source: DTI, UK energy and CO2 emissions projections, Feb 2006

6. Energy Efficiency Policies Included in Projections

7. Global GHG mitigation in context:energy in world GDP

8. Global GHG mitigation in context:energy in world GDP

9. Global GHG mitigation in context:energy in world GDP

10. Macroeconomic Modeling of Energy CGE models treatment of energy as factor of production in a production function with labor and capital Econometric models traditional approach with energy demand equations Literature on income and price elasticities

11. Macroeconomic Costs of Mitigation Costs not directly observable from market prices outcome of complex energy-environment-economy (E3) system involve changes in environment that have no market valuations hypothetical: comparison of 2 states of the E3 system over future years Macroeconomic costs usually measured in terms of future loss of GDP, comparing one hypothetical state of the world with another Debate: are such costs to be offset by ancillary benefits and benefits from use of tax or emission permit revenues? taxes/auctioned permits may incur high �political� costs free allocation of emission permits (as in phase I EU emissions trading scheme (ETS)) yields no revenues to recycle

12. Treatment of Technological Change in Cost Modelling Usual assumption in IPCC literature is of autonomous growth in energy efficiency, constant across all economies: therefore no effect on efficiency from stabilisation policies However there is good evidence that higher real prices of energy increase efficiencies (e.g. Popp, 2002; Jaffe, Newell and Stavins, 2003) costs of renewable power fall as markets develop (e.g. McDonald and Schrattenholzer, 2001) New research: modelling of endogenous technological change (bottom-up and top-down) and implications for policy action low-carbon paths as low-cost, even beneficial, global options Jaffe, A. B., R. G. Newell and R. N. Stavins 2003: Technological change and the environment. pp. 461-516 in K.-G. M�ler and J. R. Vincent (eds), Handbook of Environmental Economics, Elsevier Science B.V. McDonald, A. and L. Schrattenholzer, 2001: Learning Rates for Energy Technologies. Energy Policy 29: 255-261. Popp, D. 2002: Induced Innovation and Energy Prices. American Economic Review 92(1) 160-180.Jaffe, A. B., R. G. Newell and R. N. Stavins 2003: Technological change and the environment. pp. 461-516 in K.-G. M�ler and J. R. Vincent (eds), Handbook of Environmental Economics, Elsevier Science B.V. McDonald, A. and L. Schrattenholzer, 2001: Learning Rates for Energy Technologies. Energy Policy 29: 255-261. Popp, D. 2002: Induced Innovation and Energy Prices. American Economic Review 92(1) 160-180.

13. Induced Technological Change in Global Climate Models Method: introduce R&D and/or learning-by-doing into costs of energy technologies, so that higher real carbon prices induce change CGE models face special problems whole-economy increasing returns are incompatible with a general solution increased substitution possibilities (e.g. to renewable power or carbon-capture) are typically introduced in the only one sector (energy) economic growth remains largely given by assumption, with general technological change unaffected by the energy sector technologies An open question: can increased technological change lead to higher economic growth? Goulder Lawrence H. (2004) �Induced technological change and climate policy�, Pew Center on Global Climate Change http://www.pewclimate.org.Goulder Lawrence H. (2004) �Induced technological change and climate policy�, Pew Center on Global Climate Change http://www.pewclimate.org.

14. Question: Full Employment of Labor and �Efficient� Economy? Economy-wide studies that assume full efficiency report that the regulatory policies incur costs Parry and Williams (1999) - CGE compare 8 policy instruments to reduce CO2 Find: High costs of the energy efficiency regulations, exacerbated by tax interaction effects Smulders and Nooij (2003) - CGE analyse energy conservation on technology and economic growth Find: Policies that reduce the level of energy use unambiguously depress output levels Pizer et al. (2006) - CGE calibrated to sectoral models of the US Find: CAFE standards to be significantly more expensive than broad carbon taxes.

15. Alternative Approach A main alternative approach: detailed sectoral studies that feed into a CGE macroeconomic model Roland-Holst (2006) uses a CGE model for California to assess energy efficiency policies for CO2 reductions Find: CO2-efficiency policies can reduce transportation CO2 emissions by nearly 6% and increase Gross State Output by over 2%

16. The Approach of MDM-E3 (Multisectoral Dynamic Model � w/ Energy-Environment-Economy system) MDM is a multisectoral regional econometric model of the UK economy developed in the 1990s Equilibrium & constant returns to scale are not assumed The solutions are dynamic, integrated and consistent across the model and submodels Energy demand is derived from demand for heat & power from demand for final products No explicit production function 2-level hierarchy: aggregate energy demand equations and fuel share equations Aggregate demand affected by industrial output of user industry, household spending in total, relative prices, temperature, technical progress indicator, trends, efficiency policies

17. MDM-E3 Theory and Data Econometric, dynamic, structural, post-Keynesian based on time series and cross-section data cointegration techniques identify long-run trends in 22 sets of equations Structural: 50 industries, 13 energy users, 11 energy carriers, 51 HH categories Assumptions Social groups (not representative agents) i.e. parameters vary across sectors and regions Variable returns to scale and degrees of competition across sectors Path dependency and emphasis on �history� rather than �equilibrium� Short-term and long-run solutions With induced technological change Technological Progress Indicators (TPI) (incl. R&D) in many equations e.g. in energy-use, export, import, price, employment equations

18. Energy Submodel

19. MDM-E3 & Transport Top-down macroeconomic model Bottom-up transport system efficiency feedback to macro economy Efficiency improvements estimated offline Feedback from macro economy not incorporated in detailed transport sector

20. The UK Energy System in MDM-E3

21. MDM-E3 : Aggregate Energy-Demand Equations Autoregressive distributed lag (ARDL) model energy consumption (Et) depends on energy price (Pt), output (Yt), temperature (TEt) & lagged values: Et=a0+a1Pt+ a2Yt + a3TEt + a4Et-1 + a5Yt-1 + a6Pt-1 + a7TEt-1+et Re-parameterisation give error-correction mechanism (ECM) model: ?Et=b0+b1? Pt+ b2?Yt + b3?TEt + b4(Et-1 � b5Pt-1 � b6Yt-1- b7TEt-1) + et Augmented by time trends and/or accumulated investment to represent energy efficiency improvements ECM model distinguishes between long-term and adjustment parameters CAMBRIDGE ECONOMETRICS, UNIVERSITY OF CAMBRIDGE and POLICY STUDIES INSTITUTE (2005) Modelling the Initial Effects of the Climate Change Levy, A report submitted to HM Customs and Excise, 8 March 2005, available at www.customs.hmrc.gov.uk/channelsPortalWebApp Engle-Granger cointegration estimation methodCAMBRIDGE ECONOMETRICS, UNIVERSITY OF CAMBRIDGE and POLICY STUDIES INSTITUTE (2005) Modelling the Initial Effects of the Climate Change Levy, A report submitted to HM Customs and Excise, 8 March 2005, available at www.customs.hmrc.gov.uk/channelsPortalWebApp Engle-Granger cointegration estimation method

22. UK Transport Efficiency Policies Voluntary Agreements on vehicle CO2 emissions reductions European Commission and the European, Japanese and Korean Automobile Manufacturers Association to reduce average CO2 emissions from their new cars to 140 g/km by 2008 -2009 Targets are expected to be met via fuel saving technologies Company Car Tax Company cars are taxed on a percentage of their list price according to one of 21 CO2 emissions bands. Graduated Vehicle Excise Duty GVED - the annual vehicle tax charge � new cars placed in one of four VED rate bands according to their CO2 emissions Projected to reduce transport energy use by 3.1 mtoe and lower GHG�s by 2.3 MtC direct injection gasoline and direct injection diesel engines, engine improvements, weight reduction, reduced rolling resistance and aerodynamic improvements. the VA package policy has a net cost to society mainly because of increased congestion � with the lifetime cost of abatement being between �374 to �356 per tonne of carbon saved direct injection gasoline and direct injection diesel engines, engine improvements, weight reduction, reduced rolling resistance and aerodynamic improvements. the VA package policy has a net cost to society mainly because of increased congestion � with the lifetime cost of abatement being between �374 to �356 per tonne of carbon saved

23. Aside: Canadian Voluntary MOU Commits the Canadian automotive industry to 5.3Mt reduction in GHG emissions (CO2e) from the light duty vehicle sector by 2010 Reference case GHGs for the light duty vehicle sector in 2010 are 90.51 Mt of CO2e.

24. Rebound Effects Direct, indirect, economy-wide Three direct rebound effects More mileage driven More comfort taking (air-conditioning) Shift to larger vehicles These offset 25% of the estimated gross energy savings from the policies Indirect and economy-wide (result of model) Indirect and economy-wide rebound 7% beyond direct Total: 32%

25. Impacts of VAs on Key Macroeconomic Variables

26. Efficiency v. Fiscal Policies to Reduce Transport CO2 Emissions

27. Efficiency v. Fiscal Policies to Reduce Transport CO2 Emissions

28. Efficiency v. Fiscal Policies to Reduce Transport CO2 Emissions

29. Efficiency v. Fiscal Policies to Reduce Transport CO2 Emissions (no extra inflation over reference case) (no extra inflation over reference case)

30. Impacts of VA�s: Sensitivity Analysis (2010)

31. Discussion & Summary VAs 1.8% CO2 Reduction VAs v. fuel duties Achieving the same CO2 reduction, no recycling of revenues, no monetary responses Energy use in the transport sector goes down more with fuel duties than with the VAs, but at a cost of loss in GDP of 0.84% instead of a gain of 0.48%. Rate of duty on road fuels has to rise by 4.85% a year (real) Effects on inflation and growth is very marked Fuel duties increasing prices Employment is reduced by 0.5% VA�s & smaller fuel duties for inflation-neutrality More effective in reducing energy and emissions

32. Discussion & Summary Our approach enables a partial integration of top-down macroeconomic aspects and bottom-up energy systems We do not assume that resources are used at full economic efficiency Limitations Bottom-up energy savings and direct rebound effects had to be imposed on the model These are below the level of disaggregation currently in the model No feedbacks incorporated from the wider macroeconomic effects to the bottom-up energy savings Currently working to develop greater sectoral detail for better integration with the macro model