在做面板数据回归的过程中,hausman检验hausman fe re 和 hausman fe re,sigmaless 的结果矛盾,这个情况怎么选择?而且下面标红的文字提示的不知道是存在什么问题?希望可以得到解答,谢谢了!
. xtreg expenses wages agricultural property transferpay i.year,re
Random-effects GLS regression Number of obs = 310
Group variable: id Number of groups = 31
R-sq: within = 0.9524 Obs per group: min = 10
between = 0.9401 avg = 10.0
overall = 0.9383 max = 10
Wald chi2(13) = 5574.50
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
------------------------------------------------------------------------------
expenses | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
wages | .5917945 .040182 14.73 0.000 .5130391 .6705499
agricultural | .3604007 .0516306 6.98 0.000 .2592067 .4615948
property | 1.239058 .252673 4.90 0.000 .7438277 1.734287
transferpay | .1937242 .0953399 2.03 0.042 .0068615 .3805869
|
year |
2005 | 185.1916 58.75327 3.15 0.002 70.03729 300.3459
2006 | 262.6719 59.98991 4.38 0.000 145.0938 380.25
2007 | 308.7193 62.44749 4.94 0.000 186.3245 431.1142
2008 | 333.2574 65.25285 5.11 0.000 205.3642 461.1507
2009 | 472.8677 70.32497 6.72 0.000 335.0333 610.7021
2010 | 425.5874 78.12685 5.45 0.000 272.4616 578.7132
2011 | 593.6202 91.29191 6.50 0.000 414.6914 772.5491
2012 | 756.4751 103.0515 7.34 0.000 554.4978 958.4524
2013 | 934.2544 114.8439 8.13 0.000 709.1645 1159.344
|
_cons | 836.4422 129.2716 6.47 0.000 583.0744 1089.81
-------------+----------------------------------------------------------------
sigma_u | 343.49682
sigma_e | 221.28567
rho | .70670756 (fraction of variance due to u_i)
------------------------------------------------------------------------------
. est store re
. xtreg expenses wages agricultural property transferpay i.year,fe
Fixed-effects (within) regression Number of obs = 310
Group variable: id Number of groups = 31
R-sq: within = 0.9539 Obs per group: min = 10
between = 0.9380 avg = 10.0
overall = 0.9237 max = 10
F(13,266) = 423.25
corr(u_i, Xb) = 0.5362 Prob > F = 0.0000
------------------------------------------------------------------------------
expenses | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
wages | .4631687 .0504381 9.18 0.000 .36386 .5624774
agricultural | .2915869 .0578213 5.04 0.000 .1777414 .4054325
property | 1.517229 .2627083 5.78 0.000 .9999769 2.034481
transferpay | .1233559 .0940181 1.31 0.191 -.0617585 .3084704
|
year |
2005 | 206.8639 56.78445 3.64 0.000 95.05972 318.6681
2006 | 311.1358 58.86764 5.29 0.000 195.2299 427.0416
2007 | 379.3755 62.65369 6.06 0.000 256.0153 502.7358
2008 | 424.1124 66.65755 6.36 0.000 292.8689 555.3559
2009 | 597.0948 73.92906 8.08 0.000 451.5342 742.6554
2010 | 590.7348 85.27602 6.93 0.000 422.8329 758.6367
2011 | 818.1508 103.6544 7.89 0.000 614.0633 1022.238
2012 | 1031.151 119.362 8.64 0.000 796.1369 1266.166
2013 | 1260.974 135.5924 9.30 0.000 994.0028 1527.945
|
_cons | 1094.312 131.311 8.33 0.000 835.7709 1352.853
-------------+----------------------------------------------------------------
sigma_u | 539.45895
sigma_e | 221.28567
rho | .85597132 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(30, 266) = 26.94 Prob > F = 0.0000
. est store fe
. hausman fe re
Note: the rank of the differenced variance matrix (11) does not equal the number of coefficients being tested (13); be sure this is what you expect, or there may be problems computing the test. Examine the output of your estimators for anything unexpected and possibly consider scaling your variables so that the coefficients are on a similar scale.
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fe re Difference S.E.
-------------+----------------------------------------------------------------
wages | .4631687 .5917945 -.1286258 .0304862
agricultural | .2915869 .3604007 -.0688138 .0260304
property | 1.517229 1.239058 .2781716 .0719168
transferpay | .1233559 .1937242 -.0703683 .
2005bn.year | 206.8639 185.1916 21.67232 .
2006.year | 311.1358 262.6719 48.46386 .
2007.year | 379.3755 308.7193 70.65618 5.079033
2008.year | 424.1124 333.2574 90.85494 13.61227
2009.year | 597.0948 472.8677 124.2271 22.80145
2010.year | 590.7348 425.5874 165.1474 34.17886
2011.year | 818.1508 593.6202 224.5306 49.09197
2012.year | 1031.151 756.4751 274.6764 60.2302
2013.year | 1260.974 934.2544 326.7194 72.08461
------------------------------------------------------------------------------
b = consistent under Ho and Ha; obtained from xtreg
B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(11) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 8.66
Prob>chi2 = 0.6534
(V_b-V_B is not positive definite)
. hausman fe re,sigmaless
Note: the rank of the differenced variance matrix (4) does not equal the number of coefficients being tested (13); be sure this is what you expect, or there may be problems computing the test. Examine the output of your estimators for anything unexpected and possibly consider scaling your variables so that the coefficients are on a similar scale
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fe re Difference S.E.
-------------+----------------------------------------------------------------
wages | .4631687 .5917945 -.1286258 .03239
agricultural | .2915869 .3604007 -.0688138 .029584
property | 1.517229 1.239058 .2781716 .0995254
transferpay | .1233559 .1937242 -.0703683 .020582
2005bn.year | 206.8639 185.1916 21.67232 5.333793
2006.year | 311.1358 262.6719 48.46386 11.55071
2007.year | 379.3755 308.7193 70.65618 17.7458
2008.year | 424.1124 333.2574 90.85494 22.38237
2009.year | 597.0948 472.8677 124.2271 29.77524
2010.year | 590.7348 425.5874 165.1474 40.25818
2011.year | 818.1508 593.6202 224.5306 55.02643
2012.year | 1031.151 756.4751 274.6764 66.44546
2013.year | 1260.974 934.2544 326.7194 78.57488
------------------------------------------------------------------------------
b = consistent under Ho and Ha; obtained from xtreg
B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(4) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 27.70
Prob>chi2 = 0.0000
(V_b-V_B is not positive definite)