Dr Mona Chitnis, University of Surrey, United Kingdom
Prof Steve Sorrell, London School of Economics, United Kingdom
Dr Roger Fouquet, London School of Economics, United Kingdom
Improved energy efficiency is widely expected to play a key role in reducing energy consumption and GHG emissions. However, the energy and emissions savings from such improvements may be less than simple calculations suggest, owing to a variety of economic mechanisms that go under the heading of rebound effects. Direct rebound effects result from increased consumption of relatively cheaper energy services: for example, an efficient boiler lowers the cost of space heating so households may choose to increase internal temperatures and/or leave the heating on for longer. Indirect rebound effects result from induced changes in consumption of other goods and services, the provision of which necessarily involves energy use and GHG emissions. For example, the money saved on space heating may be spent instead on increased lighting, or on electronic goods. Re-spending therefore may lead to additional energy use and emissions, which offset the original energy and emission savings.
This study estimates the direct and indirect rebound effects following energy efficiency improvements by an average UK household. Using a unique dataset, the study separately investigates the effect of efficiency improvements for lighting, space heating, water heating, electrical appliances and cooking. This is the first study to investigate rebound effect at this level of disaggregation. The study includes the indirect rebound effects that result from increased consumption of other energy services (e.g. cheaper heating leading to more lighting), but excludes embodied energy and emissions.
This study estimates a linear Almost Ideal Demand System (AIDS) for expenditure on energy services by an average UK household. The dependent variable in each equation is the share of the individual energy service in total expenditure on energy services, while the explanatory variables are the real total expenditure on energy services, the real price of each energy service and a one period lag of the dependent variable. The data are annual time series for 1970-2013, derived from a variety of sources with estimates of average energy efficiency being used to derive the price of the individual energy services. The system of equations is estimated by econometrics approach of Iterative Seemingly Unrelated Regressions (ISUR). From this we obtain the own-price, cross-price and expenditure elasticities for each energy service. The direct rebound effect is estimated from the negative of the own- price elasticity of each energy service, while the indirect rebound effects are estimated from the cross-price elasticities and the relevant energy and GHG intensities. Rebound effects are therefore estimated in terms of both energy use and GHG emissions.
The initial results are derived assuming that total expenditure on energy services remain constant. These suggest a direct rebound effect from energy efficiency improvement of 118% for lighting, 72% for space heating, 70% for water heating, 92% for appliances and 120% for cooking. Indirect rebound effects are still to be estimated, but these are likely to increase the estimated rebound for each energy service. We are further investigating the use of a model that allows for changes in total energy service expenditure. It is expected that rebound effects would be lower when allowing for changes in total energy service expenditure.
These initial results suggest how rebound effects may vary between different energy services. Rebound effects appear to be significant for measures that improve the efficiency for space heating, water heating and electrical appliances and greater than 100% (backfire) for measures that improve the efficiency lighting and cooking. However, these results are subject to a number of caveats, and further elaborations of the model (to be incorporated into this paper) may modify these estimates.
Post your comments and questions for the speakers here