How Do Motorways Shape Commuting Patterns? An Evaluation Based On Time Series
T Nielsen, H Hovgesen, Aalborg University, DK
This paper is an offspring from a larger research project addressing the effect of the Danish motorway network on urbanization and spatial interaction patterns in the last 20 years. This paper will present results on how and to what a degree the building of the motorway network has shaped spatial interactions patterns in Denmark.
Almost half of Denmark?s 1000 km of motorway has been completed within the last 20 years. The limited access motorway, allowing speeds until 110 km/h (soon 130 km/h), is the top of the road hierarchy in most parts of the country and forms a continuous network connecting cities all over the country through high-speed corridors. Theoretically this development has a strong potential to affect the pattern of transportation as well as urban development in general. However as the debate on induced demand shows it is a difficult task to reveal an effect of road building in itself. Ordinary cross-sectional analysis will always run into difficulties when it comes to sorting out the direction of causality: was it the motorway ? or ? was it just build in response to adequate forecasts. Time series lending itself to analysis of variations in the ?response? to the motorway (before and after) as well as development trends before the road was build could produce some interesting results. Especially the inclusion of the development in the years before a road was build, as an explanatory factor in competition with motorway development ? will allow for at fairly direct testing of the claim that the motorway was merely a response to an ongoing development now producing the apparent relation between road and transportation.
The Danish commuter statistics provides a rare opportunity to analyse spatial interaction patterns over time. Statistics Denmark has kept records of place of work and place of residence at the address level, for the entire workforce since 1981. The database has been constructed and errors continuously corrected so that the register should reflect the actual commuter relations (actual place of work as opposed to company address). However no statistics is provided as to how often the commute is made ? thus only commuter ?relations? can be analysed. Given the high level of regularity of travel to work the commuter ?relations? is anticipated to reflect actual home-work travel flows adequately. The basic data used in this paper is a matrix of the exchange of commuters between some 1500 parish-derived geographical units ? by level of education (6. groupings) ? and three points in time within the last 20 years: 1982, 1992 and 2002.
To explore the effect of the completion of the motorway network on commuting patterns three types of data set are build, to make the analysis of three dependent variables possible: commuter flows between pairs of origins and destinations, commuting distances and commuting to the CBD of the largest cities. In all three types of analysis the dependent variable will be the change between 1992 and 2002.
The changing commute will be explained using multivariate regression analysis, with the change of general accessibility (travel time compared to distance-based measures), attractions, origin- and destination characteristics, travel times and access-egress to/from motorway as the main categories of variables. Finally the pace of change in the commute in the ?before? period will be included in the analysis. Thus the analysis will address the question of how the motorway construction actually changed the patterns of growth in commuter flows.
The explanatory variables mentioned above mainly describe accessibility and the general ?opportunity surface?. As a means to control for the significance of socio-economic differences for the result, the same analysis will be made for the different educational groupings.
Conclusions will be drawn as to how the motorway construction and its direct effects: increased accessibility, reduced travel time and shorter access-egress to/from the network has shaped the commuting patterns in terms of interactions between origins and destinations, commuting distances and commutes to the CBD of the largest cities. Specific notion will be given as to how the inclusion of development trends in the early period (1982-1992) as an explanatory variable effects the performance of other explanatory factors, when trying to explain the development in commuting patterns in the second period (1992-2002).
Association for European Transport