急！求问各位大侠stata面板数据共线性问题

YouHeBK

8262

收藏 2016-02-04

面板数据设置了年份、行业、地区三个虚拟变量，代码如下：

tabulate provinceid,gen(PRO)

tabulate industry_4 , gen(IND)

tabulate year,gen(REG)

drop REG1

drop IND1

drop PRO1

xtreg lresearch vsum hhilnlabor1 ros1 lsubsidy soe private foreign REG* PRO* IND*,fe

三个虚拟变量都出现共线性问题：

结果如下：

xtreg lresearch vsum hhi lnlabor1 ros1 lsubsidy soe private foreign REG* PRO* IND*,fe

note: REG2 omitted because of collinearity

note: REG3 omitted because of collinearity

note: REG6 omitted because of collinearity

note: REG7 omitted because of collinearity

note: PRO2 omitted because of collinearity

note: PRO3 omitted because of collinearity

note: PRO4 omitted because of collinearity

note: PRO5 omitted because of collinearity

note: PRO6 omitted because of collinearity

note: PRO7 omitted because of collinearity

note: PRO8 omitted because of collinearity

note: PRO9 omitted because of collinearity

note: PRO10 omitted because of collinearity

note: PRO11 omitted because of collinearity

note: PRO12 omitted because of collinearity

note: PRO13 omitted because of collinearity

note: PRO14 omitted because of collinearity

note: PRO15 omitted because of collinearity

note: PRO16 omitted because of collinearity

note: PRO17 omitted because of collinearity

note: PRO18 omitted because of collinearity

note: PRO19 omitted because of collinearity

note: PRO20 omitted because of collinearity

note: PRO21 omitted because of collinearity

note: PRO22 omitted because of collinearity

note: PRO23 omitted because of collinearity

note: PRO24 omitted because of collinearity

note: PRO25 omitted because of collinearity

note: PRO26 omitted because of collinearity

note: PRO27 omitted because of collinearity

note: PRO28 omitted because of collinearity

note: PRO29 omitted because of collinearity

note: IND2 omitted because of collinearity

note: IND3 omitted because of collinearity

note: IND4 omitted because of collinearity

note: IND5 omitted because of collinearity

Fixed-effects (within) regression Number of obs = 339

Group variable: id Number of groups = 249

R-sq: within = 0.2790 Obs per group: min = 1

between = 0.0290 avg = 1.4

overall = 0.0524 max = 3

F(10,80) = 3.10

corr(u_i, Xb) = -0.2694 Prob > F = 0.0022

------------------------------------------------------------------------------

lresearch | Coef. Std. Err. t P>|t| [95% Conf. Interval]

-------------+----------------------------------------------------------------

vsum | .0015206 .0027876 0.55 0.587 -.0040269 .0070681

hhi | .5849587 13.10891 0.04 0.965 -25.5026 26.67252

lnlabor1 | .0933548 .3189847 0.29 0.771 -.5414449 .7281546

ros1 | -.7806015 .8556584 -0.91 0.364 -2.483416 .922213

lsubsidy | -.1140379 .0848694 -1.34 0.183 -.2829333 .0548575

soe | .6553425 .629896 1.04 0.301 -.5981904 1.908876

private | .8968121 .3547705 2.53 0.013 .1907963 1.602828

foreign | 1.857314 1.202142 1.55 0.126 -.5350247 4.249654

REG2 | 0 (omitted)

REG3 | 0 (omitted)

REG4 | -1.331041 .4414639 -3.02 0.003 -2.209582 -.4524997

REG5 | -1.148099 .4552052 -2.52 0.014 -2.053987 -.2422121

REG6 | 0 (omitted)

REG7 | 0 (omitted)

PRO2 | 0 (omitted)

PRO3 | 0 (omitted)

PRO4 | 0 (omitted)

PRO5 | 0 (omitted)

PRO6 | 0 (omitted)

PRO7 | 0 (omitted)

PRO8 | 0 (omitted)

PRO9 | 0 (omitted)

PRO10 | 0 (omitted)

PRO11 | 0 (omitted)

PRO12 | 0 (omitted)

PRO13 | 0 (omitted)

PRO14 | 0 (omitted)

PRO15 | 0 (omitted)

PRO16 | 0 (omitted)

PRO17 | 0 (omitted)

PRO18 | 0 (omitted)

PRO19 | 0 (omitted)

PRO20 | 0 (omitted)

PRO21 | 0 (omitted)

PRO22 | 0 (omitted)

PRO23 | 0 (omitted)

PRO24 | 0 (omitted)

PRO25 | 0 (omitted)

PRO26 | 0 (omitted)

PRO27 | 0 (omitted)

PRO28 | 0 (omitted)

PRO29 | 0 (omitted)

IND2 | 0 (omitted)

IND3 | 0 (omitted)

IND4 | 0 (omitted)

IND5 | 0 (omitted)

_cons | 7.362842 1.906623 3.86 0.000 3.56854 11.15714

-------------+----------------------------------------------------------------

sigma_u | 2.2161597

sigma_e | .77200962

rho | .89178144 (fraction of variance due to u_i)

------------------------------------------------------------------------------

F test that all u_i=0: F(248, 80) = 6.22 Prob > F = 0.0000

求问如何解决

附件列表

gongxianxing.PNG

原图尺寸 24.28 KB

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

全部回复

liuzhen123123

2016-2-4 22:10:06

由于你选择的是FE估计，而设定的虚拟变量全部都相当于截距项发生变化，所以stata估计不出来你的虚拟变量的参数。（类似于虚拟变量与individual effect线性相关了）。

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

YouHeBK

2016-2-4 22:22:20

liuzhen123123 发表于 2016-2-4 22:10
由于你选择的是FE估计，而设定的虚拟变量全部都相当于截距项发生变化，所以stata估计不出来你的虚拟变量的参 ...

不是很明白，怎样才是正确的做法呢

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

liuzhen123123

2016-2-4 23:53:01

YouHeBK 发表于 2016-2-4 22:22
不是很明白，怎样才是正确的做法呢

用随机效应RE

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

YouHeBK

2016-2-5 10:17:00

liuzhen123123 发表于 2016-2-4 23:53
用随机效应RE

用随机效应共线性问题确实变小了，但问题依然存在，结果如下：
xtreg  lresearch vsum hhi lnlabor1 ros1 lsubsidy soe private foreign REG* PRO* IND*,re
note: REG2 omitted because of collinearity
note: REG5 omitted because of collinearity
note: REG6 omitted because of collinearity
note: REG7 omitted because of collinearity
note: PRO7 omitted because of collinearity
note: PRO21 omitted because of collinearity
note: PRO24 omitted because of collinearity
note: PRO28 omitted because of collinearity
note: PRO29 omitted because of collinearity

Random-effects GLS regression                Number of obs    =    339
Group variable: id                            Number of groups =    249

R-sq:  within  = 0.0979                      Obs per group: min =       1
   between = 0.4767                                     avg =    1.4
   overall = 0.4892                                     max =       3

                                             Wald chi2(37)    = 198.77
corr(u_i, X) = 0 (assumed)                   Prob > chi2       = 0.0000

------------------------------------------------------------------------------
lresearch |    Coef. Std. Err.    z P>|z|    [95% Conf. Interval]
-------------+----------------------------------------------------------------
      vsum | .0023162 .0025001    0.93 0.354    -.002584 .0072164
      hhi | .4577016 11.60256    0.04 0.969 -22.28289 23.19829
lnlabor1 | .7070137 .0906137    7.80 0.000    .5294141 .8846133
      ros1 | .5586112 .3730528    1.50 0.134 -.1725589 1.289781
lsubsidy | .1071422 .0472872    2.27 0.023    .0144611 .1998233
      soe | .049736 .4381083    0.11 0.910 -.8089406 .9084126
   private | .472537 .2165465    2.18 0.029    .0481136 .8969603
   foreign | .5193684 .3211996    1.62 0.106 -.1101712 1.148908
      REG2 |       0  (omitted)
      REG3 | .5866695 .3710565    1.58 0.114 -.1405879 1.313927
      REG4 |  -.0356062 .2204048 -0.16 0.872 -.4675916 .3963793
      REG5 |       0  (omitted)
      REG6 |       0  (omitted)
      REG7 |       0  (omitted)
      PRO2 |  -.6727465 .8511793 -0.79 0.429 -2.341027 .9955344
      PRO3 |  -2.075265 .9074633 -2.29 0.022    -3.85386 -.2966692
      PRO4 |  -3.220087 1.281755 -2.51 0.012 -5.732281 -.7078925
      PRO5 | 2.12665 1.776995    1.20 0.231 -1.356196 5.609496
      PRO6 |  -2.515329 1.225881 -2.05 0.040 -4.918011 -.1126462
      PRO7 |       0  (omitted)
      PRO8 |  -.7254231 1.222893 -0.59 0.553 -3.122249 1.671403
      PRO9 |  -.5656478 .5345239 -1.06 0.290 -1.613296 .4819998
   PRO10 |  -.0858753 .5785998 -0.15 0.882    -1.21991 1.048159
   PRO11 |  -1.511633 .4366466 -3.46 0.001 -2.367445 -.6558217
   PRO12 |  -1.928918 .8022319 -2.40 0.016 -3.501264 -.3565726
   PRO13 |  -1.064023 .4804511 -2.21 0.027    -2.00569 -.1223562
   PRO14 |  -.5838177 1.297846 -0.45 0.653    -3.12755 1.959914
   PRO15 |  -.2423671 1.069616 -0.23 0.821 -2.338775 1.854041
   PRO16 |  -2.673526 1.279329 -2.09 0.037 -5.180965 -.1660873
   PRO17 |  -1.201143 .9399534 -1.28 0.201 -3.043418 .6411313
   PRO18 |  -2.370076 .7735942 -3.06 0.002 -3.886292 -.8538589
   PRO19 |  -.8302714 .4065542 -2.04 0.041 -1.627103 -.0334398
   PRO20 |  -2.275143 .8350239 -2.72 0.006    -3.91176 -.6385266
   PRO21 |       0  (omitted)
   PRO22 | -1.97852 1.067705 -1.85 0.064 -4.071184 .1141429
   PRO23 |  -.9463541 .4851692 -1.95 0.051 -1.897268 .0045601
   PRO24 |       0  (omitted)
   PRO25 |  -1.086027 1.235169 -0.88 0.379 -3.506913 1.334859
   PRO26 |  -.9099553 1.244921 -0.73 0.465 -3.349955 1.530045
   PRO27 |  -1.400949 1.765517 -0.79 0.427 -4.861299 2.059401
   PRO28 |       0  (omitted)
   PRO29 |       0  (omitted)
      IND2 |  -2.693401 2.18549 -1.23 0.218 -6.976883 1.590081
      IND3 |  -.2242765 .595456 -0.38 0.706 -1.391349 .9427958
      IND4 |  -1.336956 1.935964 -0.69 0.490 -5.131376 2.457463
      IND5 |  -.6220279 .6795258 -0.92 0.360 -1.953874 .7098182
   _cons | 3.279821 .761345    4.31 0.000    1.787612    4.77203
-------------+----------------------------------------------------------------
   sigma_u |  1.4658252
   sigma_e |  .77200962
      rho | .7828501 (fraction of variance due to u_i)
------------------------------------------------------------------------------

. xtreg  lresearch vsum hhi lnlabor1 ros1 lsubsidy soe private foreign REG* PRO* IND*,re robust
note: REG2 omitted because of collinearity
note: REG5 omitted because of collinearity
note: REG6 omitted because of collinearity
note: REG7 omitted because of collinearity
note: PRO7 omitted because of collinearity
note: PRO21 omitted because of collinearity
note: PRO24 omitted because of collinearity
note: PRO28 omitted because of collinearity
note: PRO29 omitted because of collinearity

Random-effects GLS regression                Number of obs    =    339
Group variable: id                            Number of groups =    249

R-sq:  within  = 0.0979                      Obs per group: min =       1
   between = 0.4767                                     avg =    1.4
   overall = 0.4892                                     max =       3

                                             Wald chi2(36)    =       .
corr(u_i, X) = 0 (assumed)                   Prob > chi2       =       .

                                 (Std. Err. adjusted for 249 clusters in id)
------------------------------------------------------------------------------
         |             Robust
lresearch |    Coef. Std. Err.    z P>|z|    [95% Conf. Interval]
-------------+----------------------------------------------------------------
      vsum | .0023162 .0029899    0.77 0.439 -.0035439 .0081762
      hhi | .4577016 12.24461    0.04 0.970 -23.54129 24.45669
lnlabor1 | .7070137 .0975989    7.24 0.000    .5157234    .898304
      ros1 | .5586112 .6028119    0.93 0.354 -.6228785 1.740101
lsubsidy | .1071422 .0503156    2.13 0.033    .0085255 .2057589
      soe | .049736 .4451589    0.11 0.911 -.8227594 .9222313
   private | .472537 .1811693    2.61 0.009    .1174517 .8276222
   foreign | .5193684 .368236    1.41 0.158 -.2023608 1.241098
      REG2 |       0  (omitted)
      REG3 | .5866695 .568013    1.03 0.302 -.5266155 1.699954
      REG4 |  -.0356062 .2409854 -0.15 0.883 -.5079288 .4367165
      REG5 |       0  (omitted)
      REG6 |       0  (omitted)
      REG7 |       0  (omitted)
      PRO2 |  -.6727465 .9122044 -0.74 0.461 -2.460634 1.115141
      PRO3 |  -2.075265 .510092 -4.07 0.000 -3.075027 -1.075503
      PRO4 |  -3.220087 .4321022 -7.45 0.000 -4.066991 -2.373182
      PRO5 | 2.12665 .4125928    5.15 0.000    1.317983 2.935317
      PRO6 |  -2.515329 .4402904 -5.71 0.000 -3.378282 -1.652376
      PRO7 |       0  (omitted)
      PRO8 |  -.7254231 .4408533 -1.65 0.100    -1.58948 .1386335
      PRO9 |  -.5656478 .4996717 -1.13 0.258 -1.544986 .4136906
   PRO10 |  -.0858753 .5266995 -0.16 0.870 -1.118187 .9464368
   PRO11 |  -1.511633 .4578593 -3.30 0.001 -2.409021 -.6142456
   PRO12 |  -1.928918 1.09084 -1.77 0.077 -4.066926 .2090889
   PRO13 |  -1.064023 .4067766 -2.62 0.009 -1.861291 -.2667554
   PRO14 |  -.5838177 1.183184 -0.49 0.622 -2.902816 1.735181
   PRO15 |  -.2423671 1.150534 -0.21 0.833 -2.497372 2.012638
   PRO16 |  -2.673526 .7326455 -3.65 0.000 -4.109485 -1.237567
   PRO17 |  -1.201143    .98684 -1.22 0.224 -3.135314 .7330274
   PRO18 |  -2.370076 .4152416 -5.71 0.000 -3.183934 -1.556217
   PRO19 |  -.8302714 .3690541 -2.25 0.024 -1.553604 -.1069386
   PRO20 |  -2.275143 .6470113 -3.52 0.000 -3.543262 -1.007024
   PRO21 |       0  (omitted)
   PRO22 | -1.97852 1.178987 -1.68 0.093 -4.289292 .3322513
   PRO23 |  -.9463541 .5248294 -1.80 0.071 -1.975001 .0822926
   PRO24 |       0  (omitted)
   PRO25 |  -1.086027 .7809961 -1.39 0.164 -2.616752 .4446969
   PRO26 |  -.9099553 .8910542 -1.02 0.307 -2.656389 .8364789
   PRO27 |  -1.400949    .36499 -3.84 0.000 -2.116316 -.6855817
   PRO28 |       0  (omitted)
   PRO29 |       0  (omitted)
      IND2 |  -2.693401 2.029495 -1.33 0.184 -6.671138 1.284337
      IND3 |  -.2242765 .6368778 -0.35 0.725 -1.472534 1.023981
      IND4 |  -1.336956 2.286117 -0.58 0.559 -5.817663    3.14375
      IND5 |  -.6220279 .7417812 -0.84 0.402 -2.075892 .8318365
   _cons | 3.279821 .7492114    4.38 0.000    1.811394 4.748248
-------------+----------------------------------------------------------------
   sigma_u |  1.4658252
   sigma_e |  .77200962
      rho | .7828501 (fraction of variance due to u_i)
------------------------------------------------------------------------------

.

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

ifuhuo

2016-5-16 12:03:34

liuzhen123123 发表于 2016-2-4 23:53
用随机效应RE

不应该设置省份虚拟变量，因为你用固定效应，就已经是在处理个体效应了。加入省份虚拟变量还是为了控制个体效应。

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

扫码加我 拉你入群

扫码加我 拉你入群

扫码加我 拉你入群

扫码加我 拉你入群

扫码加我 拉你入群

扫码加我 拉你入群

分享

扫码加好友，拉您进群

扫码加我拉你入群

扫码加我拉你入群

扫码加我拉你入群

扫码加我拉你入群

扫码加我拉你入群

扫码加我拉你入群