Panel data model results and diagnostics for spatial autocorrelation.
Spatial scale
Minimum comparable area (MCA)
Micro-region
Meso-region
State
Method
OLS
FE
FD
SYS-GMM
OLS
FE
FD
SYS-GMM
OLS
FE
FD
SYS-GMM
OLS
FE
FD
SYS-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 ()
−0.0164
0.0244
0.0282
0.4190***
−0.0095
−0.0516
0.0104
0.2770***
0.0203
−0.1475
−0.0203
0.2643
−0.0100
−0.1947
0.00746
0.4272
(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***
0.0039
0.0050
−0.0020
0.0303***
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*
−0.0094
0.0106
(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***
0.0120
0.0173**
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)
Constant
0.3026***
0.4060***
0.0076***
0.5623***
0.2552***
0.3722***
−0.0022
0.4332***
0.2708***
0.3596***
−0.0043
0.1526
0.3463***
0.1260
0.0097
0.1064
(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 dummies
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
Observations
10,971
10,971
7,314
10,971
1,566
1,566
1,044
1,566
402
402
268
402
81
81
54
81
-squared
0.805
0.923
0.954
—
0.910
0.960
0.978
—
0.935
0.969
0.983
—
0.940
0.977
0.987
—
Adjusted -squared
0.805
0.884
0.954
—
0.909
0.940
0.978
—
0.934
0.953
0.983
—
0.934
0.961
0.985
—
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***
0.0018
0.0570
−0.1174
0.1483
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 residuals
0.4529***
++
0.4029***
0.5357***
0.4490***
0.4459***
0.4610***
0.3380***
0.2911***
0.5996***
0.0586
++
0.1000
0.1210
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.
