Two-stage procedure of building a spatial weight matrix with the consideration of economic distance
The article discusses an essential problem of spatial econometrics which is the construction of a spatial weight matrix. This matrix expresses potential spatial interactions between the researched areas and forms the basis for spatial analyses. The objective of the paper was to consider various ways of defining the a spatial weight matrix. Consideration was given to the methods of identifying adjacency between areas in the form of neighbourhood in the sense of a common border, the adjacency identified on the basis of the physical distance and on the criterion of the k number of the nearest neighbours. Also, the issue of identifying the force of impact for the previously defined neighbourhood system was taken into account. The approaches that were considered include the standardisation of matrices by rows to unity, or the identification of the force of impact based on the physical properties of areas. The outcome of the considerations is the proposed two-stage procedure of building a spatial weight matrix. A definition of economic distance was proposed, the definition which constitutes an adequate instrument for measuring the force of interaction between neighbouring areas. The introduced economic distance is an important alternative to the identification of the force of impact of economic spatial processes in relation to the distance based on the physical properties of the researched areas and in relation to the proposal of the standardisation by rows to unity.