黃河泉 发表于 2017-9-25 16:04 
1. 去除时间趋势是加 trend 项,不是时间 dummy!2. 请试试
. xtreg construction_area degree population salary residence_price residence_sales_area landtransferringfees dyear*, fe r
note: dyear1 omitted because of collinearity
note: dyear2 omitted because of collinearity
note: dyear3 omitted because of collinearity
note: dyear4 omitted because of collinearity
note: dyear5 omitted because of collinearity
note: dyear6 omitted because of collinearity
note: dyear7 omitted because of collinearity
note: dyear8 omitted because of collinearity
note: dyear9 omitted because of collinearity
note: dyear10 omitted because of collinearity
Fixed-effects (within) regression Number of obs = 350
Group variable: year Number of groups = 10
R-sq: Obs per group:
within = 0.8908 min = 35
between = 0.9640 avg = 35.0
overall = 0.8947 max = 35
F(6,9) = 2566.27
corr(u_i, Xb) = 0.3527 Prob > F = 0.0000
(Std. Err. adjusted for 10 clusters in year)
--------------------------------------------------------------------------------------
| Robust
construction_area | Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------------------+----------------------------------------------------------------
degree | .0091235 .0009079 10.05 0.000 .0070696 .0111774
population | .1887304 .0278605 6.77 0.000 .1257056 .2517552
salary | .4310071 .0600386 7.18 0.000 .2951903 .5668239
residence_price | -.0397967 .0643257 -0.62 0.551 -.1853115 .105718
residence_sales_area | .5488597 .0270065 20.32 0.000 .4877669 .6099526
landtransferringfees | .0988654 .0142641 6.93 0.000 .0665978 .131133
dyear1 | 0 (omitted)
dyear2 | 0 (omitted)
dyear3 | 0 (omitted)
dyear4 | 0 (omitted)
dyear5 | 0 (omitted)
dyear6 | 0 (omitted)
dyear7 | 0 (omitted)
dyear8 | 0 (omitted)
dyear9 | 0 (omitted)
dyear10 | 0 (omitted)
_cons | -.3585417 .2764829 -1.30 0.227 -.9839896 .2669061
---------------------+----------------------------------------------------------------
sigma_u | .06311328
sigma_e | .09674115
rho | .29854932 (fraction of variance due to u_i)
--------------------------------------------------------------------------------------
您好,结果还是这样。我用随机效应和普通的reg都可以,只有fixed effect不可以,还请麻烦您给解答一下