Global low carbon energy investment (2015) Low carbon energy R&D investment ( ) Sectoral renewable energy targets worldwide

Transcription

1 Laggard Sectors and Green Growth Claudio Baccianti 1 1 Centre for European Economic Research (ZEW) and Tilburg University 7th Atlantic Workshop on Energy and Environmental Economics A Toxa, 28 June 2016

2 In this Paper - Overview Focus on the role of innovation policies (i.e. subsidies) in decoupling fossil fuels from economic growth; Global investment in low-carbon energy technologies was $286billions in 2015, mostly targeting renewable electricity capacity with $265.8 billions (Bloomberg and UNEP 2016). Yet, electricity is less than one third of total energy demand. In an economy with multiple and complementary energy demands (i.e. electricity, heating, transport): is the distribution of low-carbon innovation across energy demands relevant for long-run decoupling?

3 In this Paper - Overview Focus on the role of innovation policies (i.e. subsidies) in decoupling fossil fuels from economic growth; Global investment in low-carbon energy technologies was $286billions in 2015, mostly targeting renewable electricity capacity with $265.8 billions (Bloomberg and UNEP 2016). Yet, electricity is less than one third of total energy demand. In an economy with multiple and complementary energy demands (i.e. electricity, heating, transport): is the distribution of low-carbon innovation across energy demands relevant for long-run decoupling?

4 Low-Carbon Innovation and Policy: the Electricity Bias Global low carbon energy investment (2015) Low carbon energy R&D investment ( ) Sectoral renewable energy targets worldwide Number of countries

5 Outline 1 Theory - Endogenous Macro Elasticity 2 Macro Elasticity: New Empirical Evidence 3 Policy Implications

6 The Model - Production Aggregate energy-using output, Y, is divided into four sectors of energy-using activities, Y j : - Electricity generation (ELE), Heating (HET), Transportation (TRA), Non-energy Use (NEU) Final output is produced with CES technology, Y = [ j J π j Y ε 1 ε j ] ε ε 1 where J = {HET,ELE,TRA,NEU}. Each sector j uses clean inputs, X C,j, and dirty inputs, X D,j, with CES technology: [ Y j = μ j X σ 1 σ C,j Working assumption: ε < 1 and σ > 1. + ( ) σ 1 1 μ j X σ D,j, ] σ σ 1

7 The Model - Production Aggregate energy-using output, Y, is divided into four sectors of energy-using activities, Y j : - Electricity generation (ELE), Heating (HET), Transportation (TRA), Non-energy Use (NEU) Final output is produced with CES technology, Y = [ j J π j Y ε 1 ε j ] ε ε 1 where J = {HET,ELE,TRA,NEU}. Each sector j uses clean inputs, X C,j, and dirty inputs, X D,j, with CES technology: [ Y j = μ j X σ 1 σ C,j Working assumption: ε < 1 and σ > 1. + ( ) σ 1 1 μ j X σ D,j, ] σ σ 1

8 The Model - Innovation Innovation in two elds of research: low-carbon and fossil fuel energy. Successful scientists obtain monopoly rights on manufacturing energy technologies m h,jit. Production of an energy input of type h {C,D} is X h,jt = L 1 α h,jt 1 0 A ν hj (1 α) h,it mh,jit α di, where A h,it is the productivity level of the intermediate good m h,jit applied to input h and α is a share parameter. Technology wedges ν hj capture dierences in sectoral absorptive capacity.

9 Decoupling Decoupling rate ˆx t = ( ˆX D,t Ŷt) Decoupling rate with uniform clean innovation policy: ˆx t = σ θ C,t  C,t The aggregate elasticity of substitution between low-carbon and fossil fuel energy, σ, is often assumed constant, but it is not.

10 Anatomy of the Macro Elasticity Given X h,t = j X h,jt, the aggregate elasticity between clean and dirty inputs is: σ t dlnx C,t /X D,t dln(mp D,t /MP C,t ) = =(1 χ t )σ + χ t ε. where χ t j (θ jt θ t ) 2 θ t (1 θ t ) ω jt is the heterogeneity index (Obereld and Raval 2015), measuring the polarization in the use of inputs across sectors. Note: θ jt is the clean cost share, θ t the economy-wide clean cost share ω jt is the sectoral expenditure share.

11 χ t σ = ε =. σ ε

12 Estimation of σ - Novel Approach Classic factor demand approach has limitations and data constraints do not allow to estimate neither a global elasticity σ,nor σ. Step 1. Innovation-side estimation of σ The average innovator on clean and fossil fuel intermediates faces the aggregate demand of energy inputs. From the model: Ct Dt = θ t 1 θ t, so that dln C D dln p D pc = σ 1. ( ) The relative protability mirrors relative aggregate expenditures on low-carbon and fossil fuel energy. Data: rm-level data on manufacturers of energy technologies (e.g. wind and gas turbines). Step 2. Retrieve σ from σ, χ and ε. Result: σ = 1.53, σ = 1.71.

13 Estimation of σ - Novel Approach Classic factor demand approach has limitations and data constraints do not allow to estimate neither a global elasticity σ,nor σ. Step 1. Innovation-side estimation of σ The average innovator on clean and fossil fuel intermediates faces the aggregate demand of energy inputs. From the model: Ct Dt = θ t 1 θ t, so that dln C D dln p D pc = σ 1. ( ) The relative protability mirrors relative aggregate expenditures on low-carbon and fossil fuel energy. Data: rm-level data on manufacturers of energy technologies (e.g. wind and gas turbines). Step 2. Retrieve σ from σ, χ and ε. Result: σ = 1.53, σ = 1.71.

14 Estimation of σ - Novel Approach Classic factor demand approach has limitations and data constraints do not allow to estimate neither a global elasticity σ,nor σ. Step 1. Innovation-side estimation of σ The average innovator on clean and fossil fuel intermediates faces the aggregate demand of energy inputs. From the model: Ct Dt = θ t 1 θ t, so that dln C D dln p D pc = σ 1. ( ) The relative protability mirrors relative aggregate expenditures on low-carbon and fossil fuel energy. Data: rm-level data on manufacturers of energy technologies (e.g. wind and gas turbines). Step 2. Retrieve σ from σ, χ and ε. Result: σ = 1.53, σ = 1.71.

15 Estimation of σ - Novel Approach Classic factor demand approach has limitations and data constraints do not allow to estimate neither a global elasticity σ,nor σ. Step 1. Innovation-side estimation of σ The average innovator on clean and fossil fuel intermediates faces the aggregate demand of energy inputs. From the model: Ct Dt = θ t 1 θ t, so that dln C D dln p D pc = σ 1. ( ) The relative protability mirrors relative aggregate expenditures on low-carbon and fossil fuel energy. Data: rm-level data on manufacturers of energy technologies (e.g. wind and gas turbines). Step 2. Retrieve σ from σ, χ and ε. Result: σ = 1.53, σ = 1.71.

16 Uniform Clean Research Subsidy Eect on X D of the redirection of TC in a growing economy: comparison between levels of σ and χ One Sector Model Multisector - Chi=.007 Multisector - Chi=.027 Multisector - Chi=.060 Multisector - Chi=.103 Multisector - Chi= Index (base year = 1) Index (base year = 1) Years One Sector Model Multisector - Chi=.007 Multisector - Chi=.027 Multisector - Chi=.060 Multisector - Chi=.103 Multisector - Chi= Years Use of dirty inputs. Estimated Use of dirty inputs. High σ = 3. σ = 1.71.

17 Conclusions Decoupling: not only the micro-level substitution matters, but also sectoral heterogeneity; New empirical strategy to estimate σ : results send a pessimistic signal; energy-saving technology is essential for the green growth transition; Intertemporal trade-o in sectoral innovation policies (in the paper): investing in leading energy technologies maximises short-term decoupling but reduces the aggregate elasticity ofsubstitution; The diusion ofelectric vehicles shifts energy demand towards electricity generation, reducing heterogeneity and making decoupling easier.

18 Thank you!