From the results reported in Tables 2 and 3, the presence of spatial dependencies is evident. The
significant presence of spatial autoregressive structures (Table 2) strongly indicate impact from
local industrial clustering on commuting behaviour. The highly significant and statistically
stronger presence of spatial Durbin processes indicate that this impact is partly measured by the
sociodemographic variables due to the proxy nature of these determinants.
Some conclusions about the relative importances of the causal model and the spatial processes
may be drawn. Comparison of the results in Tables 2 and 3 with the results from Table 1 shows
that the impact of several sociodemographic variables are misestimated when ignoring spatial
dependencies. For the impact on out-commuting, the workforce size coefficient is overestimated
in the simple specification, whereas the impact of inhabitants per household is strengthened when
the spatial Durbin process is accounted for. Regarding the in-commuting, a higher coefficient is
estimated for the number of workplaces in the spatial Durbin version, whereas the impact of
urbanization and personal incomes are adjusted downward. Finally, for the net-commuting
specification the impact of unemployment in neighbourhood municipalities is lower when
adjusting for a spatial Durbin process, whereas the impact of number of workplaces is stable.
4. Conclusions.
Following suggestions from recent literature, the present paper has estimated causal relationship
between socio-demographic workforce characteristica and commuting behaviour. These causal
models are generalized to capture the influence of unmeasured industrial clustering patterns,
using spatial Durbin- and spatial autoregressive specifications. Finally, it is shown that the
magnitudes of the impacts on commuting behaviour are seriously misestimated for several
sociodemographic criteria when the presence of these spatial patterns is ignored.
References.
Anselin, L. 1988a: Spatial econometrics: Methods and models. Kluwer Academic Publishers.
Anselin, L. 1988b: Lagrange Multiplier Test Diagnostics for Spatial Dependence and Spatial
Heterogeneity. Geographical Analysis, 20, 1-17.
Bacher, D.-L., R. Kj0ller and K. Mohr 1995: Commuting over the Sound. Institute of Local
Government Studies - Denmark, Copenhagen.
De Falleur, M. and V. Vanderville 1994: Cross-border Flows of Workers in Europe: Facts and
Determinants. Paper presented at the 34th European Regional Science Association Congress,
Groningen.
Graversen, B., H. Hummelgaard, D. Lemmich and J. Nielsen 1997: Mobility to and from Rural
Municipalities. Institute of Local Government Studies - Denmark, Copenhagen.
Hansen, C. and M. Schack 1997: Cross-border commuting in the Danish-German border region.
Paper presented at the 37th European Regional Science Association Congress, Rome.
Kristensen, K. 1997: Land og by - forbundet via vækst?[City and country - connected via
growth?] Forthcoming from Institute of Local Government Studies - Denmark, Copenhagen.