Estimating direct rebound effects for personal automotive travel in Great Britain

Panel: 4. Mobility, transport, and smart and sustainable cities

This is a peer-reviewed paper.

Authors:
Lee Stapleton, Sussex Energy Group, SPRU (Science Policy Research Unit), University of Sussex, United Kingdom
Steve Sorrell, Sussex Energy Group, SPRU (Science Policy Research Unit), University of Sussex, United Kingdom
Tim Schwanen, School of Geography and the Environment
Oxford University, United Kingdom

Abstract

Improvements in energy efficiency reduce the effective price of energy services, thereby encouraging increased consumption of those services. For example, more fuel-efficient cars encourage more car travel. These ‘direct rebound effects’ are generally neglected by policy makers, but they reduce the energy and emission savings from efficiency improvements.

There have been several studies of direct rebound effects for personal automotive travel in the US, but no study has investigated these effects in Great Britain. This study seeks to quantify the magnitude of these effects over the last 40 years. We use aggregate time-series data on transport activity, fuel consumption and other relevant variables covering the period 1970–2011 and pay careful attention to the construction, limitations and uncertainties of this data. We take vehicle kilometres travelled (VKM) as a measure of the ‘energy service’ provided by personal automotive travel and calculate the direct rebound effect from the estimated elasticity of VKM with respect to: a) road fuel prices (£/MJ); b) vehicle fuel intensity (MJ/km); and c) energy service prices (i.e. £/km).

Our approach pays particularly careful attention to methodological issues and model diagnostics. A total of 54 regression models are estimated using static, dynamic and co-integrating techniques and these models are individually evaluated against twelve different diagnostic criteria. This yields 32 statistically significant estimates of rebound effects, ranging from 10.6% to 26.8% with a mean of 18.1%. We therefore conclude that improvements in the average, on-road fuel efficiency of cars in Great Britain have stimulated increased driving that in turn has eroded around one fifth of the potential energy savings. We further find a systematic relationship between the diagnostic performance of these models and the estimated size of the rebound effect – with more statistically robust models leading to higher estimates of the rebound effect. This is an important finding that has implications for econometric modelling more generally.