Abstract:
This work applies spatial econometrics to analyze the economic growth of Rio Grande do Sul 58 statistical comparable areas between 1939 and 2001. Moran-I tests suggest that rich areas had rich neighbors, and poor ones were agglomerated on poor neighborhoods. Exploratory spatial data analysis (ESDA) indicates that high grown clusters are located on the Serra region and the low ones on the Campanha region. The standard model indicates absolute (-convergence, but it also shows spatial autocorrelation. In order to deal with it, spatial lag and error models were tested. Both performed better than the standard model, and the spatial error seems to be the best option. Tests of structural break indicate that the Campanha region, the south of the state, has a different spatial regime than the rest of the state. Again, spatial error and lag models are appropriate; and the former has the best fitting.
Ordering information: This working paper can be ordered from Secretaria da ANPEC Av. Prof. Luciano Gualberto, 908 - Prédio FEA2 São Paulo, SP 05508-900 Brazil