I am working with a dynamic panel data model, the method I am using is a System GMM, my T = 10 and my N = 136. What I am trying to prove is the effect that the participation of women in politics has on public health spending. My dependent variable is Domestic general government health expenditure (% of GDP) and as explanatory variables I have L. Domestic general government health expenditure (% of GDP), GDP per capita, population over 65, General government net lending / borrowing% GDP, Global Gender Gap Political Empowerment subindex, democracy and Urbanization. The command I use for the regression is xtdpdgmm. The problem I have is that it does not pass the Sargan test, what could I do?
Generalized method of moments estimat | ||||||
Fitting full model: | ||||||
Step 1 f(b) | 0.00004361 | |||||
Step 2 f(b) | 0.3941367 | |||||
Group variable: id | Number of obs | 1510 | ||||
Time variable: Year | Number of groups | 140 | ||||
Moment conditions: | linear | 22 | Obs per group: | |||
nonlinear | 0 | min | 1 | |||
total | 22 | avg | 10.78571 | |||
max | 12 | |||||
(Std. Err. adjusted for 140 clusters in id) | ||||||
Domestic general government health expenditure (% of GDP) | Coef. | WC-Robust Std. Err. | z | P>|z| | [95% Conf. Interval] | |
Domestic general government health expenditure (% of GDP) | 0.3717631 | 0.0553864 | 6.71 | 0.000 | 0.2632076 | 0.4803185 |
L1. | ||||||
GDP per capita | -4.25E-08 | 9.66E-08 | -0.44 | 0.660 | -2.32E-07 | 1.47E-07 |
Population65 | 0.0937279 | 0.0415555 | 2.26 | 0.024 | 0.0122806 | 0.1751751 |
General government net lending/borrowing %GDP | -0.0321863 | 0.0059574 | -5.40 | 0.000 | -0.0438626 | -0.0205101 |
Global Gender Gap Political Empowerment subindex | 0.0026242 | 0.0022592 | 1.16 | 0.245 | -0.0018037 | 0.0070521 |
democracy | 0.000138 | 0.0001077 | 1.28 | 0.2 | -0.000073 | 0.000349 |
Urbanización | 0.0105422 | 0.0069767 | 1.51 | 0.131 | -0.0031318 | 0.0242163 |
Instruments corresponding to the linear moment conditions: | ||||||
1, model(diff): | ||||||
Domestic_Health L1.Domestic_Health L2.Domestic_Health | ||||||
2, model(diff): | ||||||
D.GDP_PPP L1.D.GDP_PPP L2.D.GDP_PPP D.Population65 L1.D.Population65 | ||||||
L2.D.Population65 D.LendingBorrowing L1.D.LendingBorrowing | ||||||
L2.D.LendingBorrowing D.GPE_subindex L1.D.GPE_subindex L2.D.GPE_subindex | ||||||
D.Polity2 L1.D.Polity2 L2.D.Polity2 D.Urbanization L1.D.Urbanization | ||||||
L2.D.Urbanization | ||||||
3, model(level): | ||||||
_cons |
. estat serial
Arellano-Bond test for autocorrelation of the first-differenced residuals
H0: no autocorrelation of order 1: z = -4.6007 Prob > |z| = 0.0000
H0: no autocorrelation of order 2: z = 0.0692 Prob > |z| = 0.9449
. estat overid
Sargan-Hansen test of the overidentifying restrictions
H0: overidentifying restrictions are valid
2-step moment functions, 2-step weighting matrix chi2(14) = 55.1791
Prob > chi2 = 0.0000
2-step moment functions, 3-step weighting matrix chi2(14) = 58.1095
Prob > chi2 = 0.0000
0 Response to Overidentification problem in a dynamic panel
Post a Comment