Rebound Effects in Europe

Rebound Effects in Europe Elena Verdolini, Maurizio Malpede V International Academic Symposium Challenges for the Energy Sector 07 February 2017, Barcelona Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 1 / 20

Outline Outline of the Presentation Motivation Objective of the Study Methodology Data Results Conclusions and Further Research Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 2 / 20

Motivation Motivation Energy efficiency improvements allow to produce the same output with less energy input. A static and simplistic expectation is that of engineering savings: X% improvements in (technical) energy efficiency X% decrease in energy demand This relationship does not hold empirically figure Output cannot be assumed constant over time Even if output were constant, energy efficiency improvements change relative prices and give rise to adjustments in production and consumption Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 3 / 20

Motivation Economic Effects of Energy Efficiency Shocks Income effect Due to decrease in relative price of using energy, more income available to purchase other goods Substitution effect Energy (and energy intensive goods) are relatively less expensive than other (non-energy intensive) goods Dynamic Efficiency i.e. induced technical change Rebound Effect Measures the percentage of potential engineering energy savings that were not realized following an efficiency improvement due to the forces described above (Saunders, 1992). Backfire Indicates a case where energy consumption actually increases following an the energy efficiency improvement. Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 4 / 20

Motivation Why should we be concerned with Rebound Effects? Rebound effects are troublesome because they imply that energy-efficiency policies are ineffective in reducing energy demand, and associated carbon emissions. Little agreement on (1) their existence and (2) their magnitude. This is due to difficulties in measuring such effects. Different types of rebound effects can be studied: direct, indirect, economy wide. Here we focus on direct rebound effects. Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 5 / 20

Objective of the Study Objective of the Study We compute the estimated direct rebound effect for 5 countries over the years 1995-2009. R e : Percentage of engineering energy efficiency savings which are eroded due to income and subsitution effects where: R e = ED Actual Energy Demand ED ED ED ED (1) ED Energy demand with no efficiency improvements ED Energy demand under full engineering savings Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 6 / 20

Objective of the Study Objective of the Study This study attempts to quantify the R e by computing the three scenarios and backcasting different energy demands levels Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 7 / 20

Methodology Methodology Modified Translog Cost Function. It includes efficiency gains parameters characterizing how each vintage increases input efficiency over time (Saunders 2013): ln c = ln c 0 aλ T t +a ln p T + 1 2 ln p B ln pt λ B ln p T t + 1 2 λ B λt t 2 (2) where p is the vector of factor prices; B is the core Hessian matrix and λ is a vector of technology gains λ = (λ K, λ L, λ E, λ M ) Four production factors: K, L, E, M Efficiency gains parameters are estimated by using historical data Backcast different energy demands levels for each sector, by switching on and off the efficiency gains parameters (and constraining output levels) Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 8 / 20

Methodology Methodology - Computation of the 3 scenarios ED ED R e = ED ED 100% Rebound Scenario: Energy efficiency gains do not lead to any reduction in energy demand (i.e. demand is the same as with no energy efficiency gains). Simulation: λ E = 0 and back-cast ED we obtain ED. Actual energy scenario 0% Rebound Scenario: Full engineering savings equivalent to assuming Leontieff production function, which leaves economic output, output costs and other factor uses equal to those of the 100% rebound case. Simulation: Introducing measured energy technology gains to the 100% rebound case, but forcing one-for-one decreases in the energy-using capacity of each new vintage of capital Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 9 / 20

Data Data 5 European countries: Germany, Uk, France, Italy, Spain table 30 sectors 1 1995-2009 Main data source: WIOD database for output and inputs quantities Input prices, energy use, depreciations rates kindly provided by K. Kratena 1 Excluded: refined petroleum and nuclear fuel, wholesale and commission and retail trade, household consumption Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 10 / 20

Results Results Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 11 / 20

Results Results Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 12 / 20

Results Energy Only VS All Factors The rebound estimates presented so far are based on counterfactual energy demand assuming that energy technology gains did not happen (i.e.λ E = 0), but allows efficiency gains in all other factors Efficiency gains in other factors of production (i.e.λ K, λ L, λ M ) may be due to energy efficiency improvements Therefore we repeat the back-casting to estimate rebound effects in case technology gains for all factors were equal to zero (i.e.λ K = λ L = λ E = λ M = 0 ) This is an upper bound estimate of rebound. The result is consistent back-fire (rebound higher than 100%) Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 13 / 20

Results Results Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 14 / 20

Results Rebound and Carbon Emissions Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 15 / 20

Conclusions Conclusions and Further Research Energy efficiency policy alone will not lead to a reduction in energy demand. About half of the engineering savings will not materialize due to induced changes in production processes. Rebound effects are very heterogenous among the countries in our sample: inportance of sectoral composition Rebound analysis needs to be linked to the issue of (counterfactual) emissions We confirm that a successfull approach to reduce emissions needs to support carbon-free options alongside energy efficiency Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 16 / 20

Conclusions and Further Research Caveats and Further Works Econometric estimation does not perform well in all sectors. Increasing the time coverage would be crucial to improve fit of the model. The simulations only include direct rebound effects (income and substitution effects). No consideration for indirect rebound (general price effect). This means that we are likely underestimating rebound effects. Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 17 / 20

Thanks Thank you! Comments welcome elena.verdolini@feem.it Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 18 / 20

Thanks Appendix return Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 19 / 20

Thanks Appendix return Elena Verdolini, Maurizio Malpede (FEEM) Rebound Effects in Europe 07 February 2017, Barcelona 20 / 20