Estimating car use rebound effects from Swedish microdata

The direct rebound effect for private car transport was estimated by following a large sample of Swedish households (28,876) that acquired a new car in 2009. For some households, this resulted in an improvement in fuel efficiency, whereas others acquired a less or similarly fuel efficient car. The households’ travel distances were measured and analysed for a period of 3 years before and 3 years after the car was replaced. This approach differs from previous econometric analyses in which fleet-average changes in distance travelled were studied, often using fluctuations in fuel cost as a proxy for changes in fuel efficiency. No significant bivariate relationship was found between changes in fuel efficiency and annual distance travelled but a multivariate analysis that also included changes in income, number of cars in the household, car weight and car power, resulted in a significant rebound effect of 24 %. Households who bought a car that was labelled ‘green’ did not exhibit any rebound effect, while households who bought a ‘normal’ car displayed a rebound effect of 32 %. This could indicate that households that buy a car with improved fuel efficiency for environmental reasons also avoid the economically induced rebound effect. The analysis did not indicate any significant differences in the rebound effect between different socio-demographic groups.