Abstract
We apply the geographically weighted regression to investigate spatial variability (nonstationarity) in the relationships between municipal operating expenditure (in total and for a separate component thereof) and the determinants of this expenditure, in South Africa. The empirical findings indicate that some of these relationships are spatially varying for the period under consideration. The global model (i.e., least square regression) cannot take account of the unequal environment in which South African municipalities operate. One implication of our findings is that a “one-size-fits-all” approach in the design of policies targeting municipal finances may not be appropriate for municipalities in South Africa, and indeed in other contexts in which such heterogeneity is found. Instead, policy formulation should explicitly consider relevant differences in local conditions.
Original language | English |
---|---|
Pages (from-to) | 396-420 |
Number of pages | 25 |
Journal | Journal of Regional Science |
Volume | 60 |
Issue number | 2 |
DOIs | |
Publication status | Published - 1 Mar 2020 |
Keywords
- GWR
- global model
- municipal finance
- municipalities
ASJC Scopus subject areas
- Development
- Environmental Science (miscellaneous)