Bengt Kriström and Thomas Laitila
Thomas Laitila, Marie Lundgren and Michael Olsson
This chapter studies individual commuting behaviour by considering the spatial variation of the economic milieu. The model employed relates the relative commuting probabilities to the relative labour demand and relative worker competition. The authors work with aggregated data separated into categories defined by gender and education level. The dependent variables in the model are heteroskedastic and dependent. Moreover, the variables are logarithmic ratios, which is an additional statistical challenge. In order to estimate the model in this situation, the authors use Feasible generalized least squares. The model is estimated using the whole data set, but also for the categories. Overall, the estimated model fits the data well. An individual loses utility when commuting from the home municipality. The loss is greater if the commute is to another region. The individual utility is positively related to labour demand and negatively related to worker competition. This is the case for all categories, with one minor exception.