Table 2: Panel data model results and diagnostics for spatial autocorrelation.

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

Explanatory variables
 ln (income per capita)−0.0290***−0.1051***−0.1250***−0.0681***−0.0305***−0.0851***−0.1084***−0.0348***−0.0346***−0.0779***−0.1072***−0.0327***−0.0487***−0.0931***−0.1135***−0.0480
(0.0007)(0.0010)(0.0010)(0.0101)(0.0016)(0.0028)(0.0027)(0.0078)(0.0032)(0.0057)(0.0057)(0.0126)(0.0084)(0.0120)(0.0127)(0.0300)
 Population growth ( )−0.01640.02440.02820.4190***−0.0095−0.05160.01040.2770***0.0203−0.1475−0.02030.2643−0.0100−0.19470.007460.4272
(0.0112)(0.0180)(0.0180)(0.1099)(0.0211)(0.0394)(0.0366)(0.0990)(0.0432)(0.0918)(0.0897)(0.1635)(0.0893)(0.1471)(0.1548)(0.2833)
 ln (years of schooling)0.0091***0.0009**0.0015***0.0122***0.0263***0.00390.0050−0.00200.0303***0.0124**0.0166***0.0233*0.0457***0.0242*0.0314**0.0686**
(0.0004)(0.0004)(0.0005)(0.0040)(0.0015)(0.0028)(0.0031)(0.0070)(0.0032)(0.0056)(0.0064)(0.0119)(0.0095)(0.0134)(0.0159)(0.0322)
 ln (population density)−0.0006***−0.0089***−0.0098***−0.0066***−0.0013***−0.0132***−0.0121***−0.0054−0.0011**−0.0183***−0.0155***0.0094−0.0021*−0.0141*−0.00940.0106
(0.0002)(0.0013)(0.0015)(0.0015)(0.0003)(0.0024)(0.0027)(0.0040)(0.0005)(0.0047)(0.0053)(0.0063)(0.0011)(0.0084)(0.0098)(0.0089)
 ln (transportation cost to SP)−0.0129***0.0056*0.0031−0.0307***−0.0067***0.0106**0.0151***−0.0281***−0.0071***0.01200.0173**0.0015−0.0101***0.0468***0.0476***0.0096
(0.0004)(0.0028)(0.0028)(0.0059)(0.0006)(0.0050)(0.0046)(0.0068)(0.0011)(0.0084)(0.0077)(0.0109)(0.0028)(0.0161)(0.0152)(0.0167)
 Constant0.3026***0.4060***0.0076***0.5623***0.2552***0.3722***−0.00220.4332***0.2708***0.3596***−0.00430.15260.3463***0.12600.00970.1064
(0.0052)(0.0226)(0.0014)(0.0750)(0.0091)(0.0418)(0.0030)(0.0778)(0.0167)(0.0729)(0.0058)(0.1304)(0.0443)(0.1375)(0.0122)(0.2439)
 Time dummiesyesyesyesyesyesyesyesyesyesyesyesyesyesyesyesyes

Observations10,97110,9717,31410,9711,5661,5661,0441,56640240226840281815481
-squared0.8050.9230.9540.9100.9600.9780.9350.9690.9830.9400.9770.987
Adjusted -squared0.8050.8840.9540.9090.9400.9780.9340.9530.9830.9340.9610.985

Diagnostics for spatial autocorrelation:
 Moran's in the residuals (70s) 0.2737***0.3411***0.4541***0.3640***0.3558***0.4807***0.2630***01201**0.4354***0.1439−0.1471−0.0366
 Moran's in the residuals (80s) 0.2704***0.1555***0.2816***0.4001***0.3665***0.3388***0.3330***0.3668***0.1436***0.2358***0.0847*0.5517***0.00180.0570−0.11740.1483
 Moran's in the residuals (90s) 0.2426***0.3189***0.2656***0.3992***0.3978***0.4231***0.4358***0.4385***0.4778***0.4166***0.4964***0.5328***0.3306**0.2397**0.2016*0.3197**
 Moran's in the averaged residuals0.4529***++0.4029***0.5357***0.4490*** 0.4459***0.4610***0.3380*** 0.2911***0.5996***0.0586++0.10000.1210

Note: standard errors in parentheses; ***significant at 1%; **significant at 5%; *significant at 10%. Moran’s significance test based on the permutation approach with ten thousand permutations and the spatial weight matrix is the queen contiguity matrix. ++ For fixed-effects (FE) estimations, the time-averaged residuals are zero by construction; therefore Moran’s statistics are not calculated. All SYS-GMM results are from the one-step estimates and include 11 instruments. In all SYS-GMM estimations, all growth determinants are treated as potentially endogenous.