Urban developments and daily travel distances: Fixed, random and hybrid effects models using a Dutch pseudo-panel over three decades

Abstract

As people require time to adjust their travel behaviour to changes in residential location and transport infrastructure, there is a need for long-term empirical studies quantifying the relationships between locations, individuals and travel behaviour. Such empirical evidence is critical for assessing previous and candidate future land use-transport policies. Existing research however, has mostly investigated travel behaviour during relatively short time periods and for a single transport mode. This paper examines the development of travel behaviour and its socio-demographic and location determinants, using Dutch National Travel Survey data from 1980 to 2010 among other sources, for the Randstad, the Netherlands. A pseudo panel analysis is conducted to investigate the effect of various indicators on average daily distance travelled by train, car and bicycle over three decades. Econometric models including pooled ordinary least squares, fixed and random effects and a hybrid model were tested to identify the best fit. The results indicate that average daily distance travelled rose until the mid-1990s before witnessing a decrease till 2010. Interestingly, half of the Randstad inhabitants have been travelling ≤26 km per day over the past thirty years. Furthermore, as people grow older, they increasingly travel more by train and bicycle. Finally, a rise in suburban inhabitants decreases the average distance travelled by train and increases that of bicycle, while a rise in rural inhabitants encourages higher distances travelled by car.