Local concentrations
Most global indices of concentrations are obtained as weighted averages of convex functions of distributions ratios, such as the per capita income. We seek to define local indices of concentrations, comparing the wealth of a region to its neighbours, where the spatial weights defining neighbourhood are formally equivalent to the components of a reversible Markov transition matrix. Second-order local concentrations are shown to generalize Moran or Geary autocorrelation indices, while first-order local concentrations can be constructed so as to not exceed their ordinary or global counterpart. Behaviour under aggregation and the Pigou-Dalton principle are further discussed within the proposed formalism, which is exemplified on wealth distribution among the Swiss cantons under the neighbourhood structure induced by inter-regional migrations. Copyright (c) 2008 the author(s). Journal compilation (c) 2008 RSAI.
Year of publication: |
2008
|
---|---|
Authors: | Bavaud, François |
Published in: |
Papers in Regional Science. - Wiley Blackwell. - Vol. 87.2008, 3, p. 357-370
|
Publisher: |
Wiley Blackwell |
Saved in:
freely available
Saved in favorites
Similar items by person
-
Bavaud, François, (2008)
-
Testing spatial autocorrelation in weighted networks : the modes permutation test
Bavaud, François, (2016)
-
On the Schoenberg Transformations in Data Analysis: Theory and Illustrations
Bavaud, François, (2011)
- More ...