Table 9 DKSE, FE, Two-step GMM, and Granger causality results

Model 1A using DKSE when the dependent variable is LNLEXP

 PRUND

-0.00348***

0.0008573

-4.06

0.001

-0.005289

-0.0016715

F(5,17) = 435.58

Prob > F = 0.0000 within R² = 0.6434

 GDPPC

-4.28e-06

2.95e-06

-1.45

0.165

-0.0000105

1.95e-06

 GOVEXP

0.00474

0.0049357

0.96

0.350

-0.0056726

0.0151541

 MNSCHOOL

0.0936***

0.004993

18.76

0.000

0.0831156

0.1041842

 URBAN

-0.00011

0.0001784

-0.65

0.526

-0.0004919

0.0002609

 _CONS

3.7019***

0.0371405

99.67

0.000

3.623589

3.780308

Model 1B using DKSE when the dependent variable is LNLEXP

 AVRDES

0.00317***

0.000845

3.76

0.002

0.0013936

0.0049593

F(5,17) = 410.16

Prob > F = 0.0000 within R² = 0.6295

 GDPPC

-3.81e-06

2.78e-06

-1.37

0.188

-0.0000097

0.0000021

 GOVEXP

0.00188

0.00573

0.33

0.746

-0.0102006

0.0139790

 MNSCHOOL

0.0926***

0.005319

17.42

0.000

0.0814219

0.1038642

 URBAN

-0.00011

0.000166

-0.67

0.512

-0.0004620

0.0002392

 _CONS

3.2941***

0.074949

43.95

0.000

3.1360170

3.4522750

Model 1C using the FE model when the dependent variable is LNINFMOR

 PRUND

0.0119***

0.0009722

12.29

0.000

0.010040

0.013860

sigma_u = 0.45608378

sigma_e = 0.10751418

F test that all u_i = 0: F(30, 522) = 164.49 Prob > F = 0.0000

within R² = 0.6823

 GDPPC

0.000057***

0.0000087

6.53

0.000

0.000040

0.000074

 GOVEXP

-0.0103

0.0088257

-1.17

0.241

-0.027691

0.006986

 MNSCHOOL

-0.2812***

0.0110858

-25.37

0.000

-0.303052

-0.259495

 URBAN

0.00061

0.0006116

1.00

0.316

-0.000588

0.001815

 _CONS

4.9049***

0.0575012

85.30

0.000

4.791987

5.017912

Model 1D using GMM when the dependent variable is LNINFMOR

 L_LNINFMOR

32.1844***

10.5261

3.06

0.002

11.55363

52.81518

No of instruments = 21, No of groups = 31

AR(1): z = -0.98 Pr > z = 0.326

AR(2): z = -1.02 Pr > z = 0.308

Sargan test of over-identification: Chi2(15) = 4.15 Prob > chi2 = 0.997

Hansen test of over-identification: Chi2(15) = 12.22 Prob > chi2 = 0.662

 AVRDES

-0.0139***

0.0052795

-2.63

0.008

-0.024255

-0.00356

 GDPPC

0.000065

0.000082

0.8

0.424

-0.0000952

0.000226

 GOVEXP

-0.0891**

0.0372889

-2.39

0.017

-0.1622396

-0.01607

 MNSCHOOL

-0.1152

0.1239816

-0.93

0.352

-0.3582814

0.127717

 URBAN

0.0010

0.0013938

0.75

0.454

-0.0016875

0.003776

Dumitrescu and Hurlin [124] Granger causality test

 Null hypothesis

W-bar

Z-bar

Z-bar tilde

p-value

Decision

 PRUND does not Granger-cause LNLEXP

17.5322

65.0874

48.4371

0.0000

PRUND does Granger-cause LNLEXP

 AVRDES does not Granger-cause LNLEXP

13.2728

48.3179

35.8304

0.0000

AVRDES does Granger-cause LNLEXP

 PRUND does not Granger-cause LNINFMOR

5.0078

15.7787

11.3686

0.0000

PRUND does Granger-cause LNINFMOR

 AVRDES does not Granger-cause LNINFMOR

5.2776

16.8409

12.1671

0.0000

AVRDES does Granger-cause LNINFMOR