Industrial Cores and Peripheries in Brazil

Table 8: Imports (SAR error model - GM estimation)

*significant, 10%; **significant, 5%; ***significant, 1%

The results also show that municipal exporting areas are explained by the size of their
municipalities (a positive coefficient for the population variable) and also by the qualification of their
labor force (a positive coefficient for E25). Another significant variable for the volume of municipal
exports is the concentration of capital and durable consumer goods sectors. On the other hand, given
Brazil’s trend toward exporting industrial commodities, the intermediate goods sector was expected to
make a positive and significant contribution. However, a substantial portion of these sectors is geared
to the domestic market and intensive in natural resources, particularly the chemical, metal and non-
metal industries, which could explain the lack of statistical significance.

The specification tests have shown that the SARSAR (or SARMA) model is the most appropriate
for industrial exports and have strongly indicated the non-inclusion of spatially autocorrelated variables
in the import model, such as tradable commodities in which Brazil is not self-sufficient (e.g. oil).
Importing areas are those with high concentration of type A and low concentration of type C
companies, which seem to have a very low proclivity to import. As expected, the scale of the local
market, measured based on its population, has a positive effect on industrial imports. With respect to
the mix of the production structure, imports are higher where a concentration exists of durable

22

<<   <   19   20   21   22   23   24   25   >   >>