Table 6: Diagnostics for spatial autocorrelation by means of Moran’s in the time-averaged residuals of the estimations of Table 2.

Spatial scaleMinimum comparable area (MCA)Micro-regionMeso-regionState
MethodOLSFEFDSYS-GMMOLSFEFDSYS-GMMOLSFEFDSYS-GMMOLSFEFDSYS-GMM

Diagnostics for spatial autocorrelation—Spatial weight matrix:
 400 km cut-off0.1983***++0.1621***0.2589***++++++
 500 km cut-off0.1586***++0.1331***0.2184***0.1401***++0.2577***0.1653***++++
 600 km cut-off0.1262***++0.1076***0.1852***0.1156***++0.2134***0.1274***0.1742***++0.2561***0.4874***++
 700 km cut-off0.1010***++0.0874***0.1614***0.0842***++0.1812***0.0835***0.1373***++0.2070***0.4484***++
 800 km cut-off0.0810***++0.0685***0.1416***0.0601***++0.1488***0.0487***0.0950***++0.2224***0.4130***0.1718*++0.01580.1334
 900 km cut-off0.0675***++0.0524***0.1253***0.0448***++0.1292***0.0239***0.0785***++0.1493***0.3990***0.0947++0.07650.1026
 1,000 km cut-off0.0600***++0.0409***0.1148***0.0387***++0.1104***0.0080**0.0707***++0.1258***0.3623***0.0561++0.01600.0481
 1,500 km cut-off0.0459***++0.0274***0.0784***0.0318***++0.0687***−0.00530.0304***++0.0789***0.2300***−0.0573++0.0358−0.0594
 2,000 km cut-off0.0236***++0.0169***0.0417***0.0106***++0.0441***0.0053*−0.0029++0.0480***0.1267***−0.0783++0.00980.0286

Note: ***significant at 1%, **significant at 5%, *significant at 10% based on the permutation approach with ten thousand permutations. The residuals of the panel data estimations were time averaged and then Moran’s statistics were calculated. ++ For fixed-effects (FE) estimations, the time-averaged residuals are zero by construction; therefore, Moran’s is not calculated.