我用jplv7的demopanelscompare程序修改成自己的数据跑,结果R^2出现了负值,想问问大神是怎么回事?
Pooled model with spatially lagged dependent variable and spatial fixed effects
Dependent Variable = gy
R-squared = 0.5823
corr-squared = 0.0869
sigma^2 = 0.0001
Nobs,Nvar,#FE = 310, 2, 32
log-likelihood = 1050.909
# of iterations = 1
min and max rho = -1.0000, 1.0000
total time in secs = 0.2480
time for optimiz = 0.0820
time for lndet = 0.0330
time for t-stats = 0.0070
No lndet approximation used
***************************************************************
Variable Coefficient Asymptot t-stat z-probability
y0 -0.013477 -4.252976 0.000021
W*dep.var. 0.614994 12.559228 0.000000
LR-test joint significance spatial fixed effects, degrees of freedom and probability = 154.6733, 31, 0.0000
Pooled model with spatially lagged dependent variable and spatial random effects
Dependent Variable = gy
R-squared = 0.5236
corr-squared = 0.0003
sigma^2 = 0.0001
Nobs,Nvar = 310, 3
log-likelihood = 1000.0818
# of iterations = 100
min and max rho = -1.0000, 1.0000
total time in secs = 2.3760
time for optimiz = 2.3580
time for lndet = 0.0090
time for t-stats = 0.0050
No lndet approximation used
***************************************************************
Variable Coefficient Asymptot t-stat z-probability
intercept 0.053065 3.166297 0.001544
y0 -0.004848 -2.252577 0.024286
W*dep.var. 0.626965 13.082215 0.000000
teta 0.384518 5.886482 0.000000
LR-test significance spatial random effects, degrees of freedom and probability = 53.0189, 1, 0.0000
Hausman test-statistic, degrees of freedom and probability = -1.3745, 2, 0.5029
Pooled model with spatial error autocorrelation and spatial fixed effects
Dependent Variable = gy
R-squared = -2.8746
corr-squared = 0.1226
sigma^2 = 0.0000
log-likelihood = 1115.5494
Nobs,Nvar,#FE = 310, 1, 32
# iterations = 19
min and max rho = -0.9900, 0.9900
total time in secs = 0.2070
time for optimiz = 0.1720
time for lndet = 0.0100
time for t-stats = 0.0020
No lndet approximation used
***************************************************************
Variable Coefficient Asymptot t-stat z-probability
y0 -0.154655 -17.186746 0.000000
spat.aut. 0.935977 72.488020 0.000000
LR-test joint significance spatial fixed effects, degrees of freedom and probability = 283.9540, 31, 0.0000
Pooled model with spatial error autocorrelation and spatial random effects
Dependent Variable = gy
R-squared = 0.5130
corr-squared = 0.0015
sigma^2 = 0.0001
Nobs,Nvar = 310, 2
log-likelihood = 1002.4364
# of iterations = 10
min and max rho = -1.6126, 1.0000
total time in secs = 0.2290
time for optimiz = 0.2140
time for eigs = 0.0020
time for t-stats = 0.0040
***************************************************************
Variable Coefficient Asymptot t-stat z-probability
intercept 0.052309 2.717294 0.006582
y0 -0.000944 -0.378272 0.705228
spat.aut. 0.650770 13.889468 0.000000
teta 0.330900 3.126890 0.001767
LR-test significance spatial random effects, degrees of freedom and probability = 57.7280, 1, 0.0000
Hausman test-statistic, degrees of freedom and probability = -275.9955, 2, 0.0000