The rebound effect is the causal impact of an increase in fuel efficiency on the demand for car kilometres. The rebound effect has been studied in several different sectors (for a survey, see Greening et al. (2000)). Some recent studies have extensively focused on passenger transportation using macro data (see, Small and Van Dender (2007); Hymel et al. (2010)). However, the rebound effect is essentially a phenomenon that occurs at the level of the individual actor and it seems appropriate to investigate it on micro data as well. This is the purpose of the present paper. We estimate the rebound effect using individual Danish administrative register data. Central to the analysis is a demand equation for car kilometres in which the fuel price and fuel efficiency are the main explanatory variables. We start by imposing the restriction that only fuel cost per kilometre matter (as is suggested by conventional theory) but also estimate versions in which the coefficients for fuel price and fuel efficiency can differ. We use administrative register micro data over the period 2004- 2010 to estimate the demand equation. We apply fixed-effect panel-data (first-diff.) techniques. The focus is on car users that switch cars during the period of observation. Endogeneity of the fuel efficiency of the new car is an important concern. We deal with endogeneity of car characteristics, following Berry, Levinsohn and Pakes (1995), by instrumenting them using the characteristic of the old car relative to the average Danish car. We also model the car switching decision using a Heckman model and incorporating the age of the old car as an additional variable. Preliminary results include the following. First, the data reject the conventional formulation in which only fuel cost per kilometre matters. If we include fuel price and fuel efficiency separately in the equation, the coefficients differ. Second, the Heckman selection equation confirms higher fuel prices induce households to switch car, probably a more fuel efficient one. Third, although preliminary, the results suggest the presence of a rebound effect of the same order of magnitude that has been found in the macro studies.
Main Research Area:
Annual Conference of the International Transportation Economics Association, 2014