deem 发表于 2017-8-5 10:27 
如果你的日期是数字1,2,3,。。。这样就没有影响,即间隔为1的数字。如果不是,请转为间隔为1的数字
那我的模型存在组间异方差和组内自相关,所以我进行FGLS时候怎么提示如下内容
xtgls pl car1 lnla1 roa1 ld1 nil1 ea1 dl1 bb1 gdp1 cpi1 id2-id16 time,p(h) corr(ar1)
time is not regularly spaced or does not have intervals of delta -- use
the force option to treat the intervals as though they were regular
以下为组间异方差修正和组内自相关检验结果,谢谢您的解答。
. xtgls pl car1 lnla1 roa1 ld1 nil1 ea1 dl1 bb1 gdp1 cpi1,p(h)
Cross-sectional time-series FGLS regression
Coefficients: generalized least squares
Panels: heteroskedastic
Correlation: no autocorrelation
Estimated covariances = 16 Number of obs = 169
Estimated autocorrelations = 0 Number of groups = 16
Estimated coefficients = 11 Obs per group: min = 8
avg = 10.5625
max = 11
Wald chi2(10) = 236.09
Prob > chi2 = 0.0000
------------------------------------------------------------------------------
pl | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
car1 | -.7851955 .3751738 -2.09 0.036 -1.520523 -.0498683
lnla1 | 1.762098 .3179013 5.54 0.000 1.139023 2.385173
roa1 | 7.637742 2.408147 3.17 0.002 2.91786 12.35762
ld1 | 9.795191 6.229347 1.57 0.116 -2.414105 22.00449
nil1 | 48.41679 23.77969 2.04 0.042 1.809456 95.02412
ea1 | -.2216278 .5471594 -0.41 0.685 -1.294041 .8507849
dl1 | -1.762921 1.477234 -1.19 0.233 -4.658247 1.132404
bb1 | -.0028928 .0031723 -0.91 0.362 -.0091105 .0033249
gdp1 | -.8132414 .2626436 -3.10 0.002 -1.328014 -.2984694
cpi1 | .2725489 .1948539 1.40 0.162 -.1093578 .6544557
_cons | 8.70783 8.861158 0.98 0.326 -8.65972 26.07538
------------------------------------------------------------------------------
. xtserial pl car1 lnla1 roa1 ld1 nil1 ea1 dl1 bb1 gdp1 cpi1 id2-id16 time
Wooldridge test for autocorrelation in panel data
H0: no first-order autocorrelation
F( 1, 15) = 28.867
Prob > F = 0.0001