Excess commuting and the modifiable areal unit problem

Excess commuting has emerged during the past two decades as an important construct for evaluating the spatial relationships between employment and residential locations. During this time-period, there has been an on-going debate regarding how one should measure excess commuting in urban regions. This paper contributes to the debate by focusing on the spatial issues inherent in excess commuting evaluation. We demonstrate how scale and unit definition (the modifiable areal unit problem) are manifested in the assessment of excess commuting, both in theoretical and empirical terms. To this end, a geographical information systems-based analysis is presented which explores spatial sensitivities in the excess commuting measure. Our results show that aggregation and spatial unit definition may have profound impacts on the estimation of excess commuting. This work provides a formal resolution to much of the recent debate regarding estimates of excess commuting in urban regions.