求助：面板数据实证问题

filippofox

2386

收藏 2007-05-07

有如下分析结果，关键解释变量B1 B2 B3，控制变量A1-A6。

目的是求得解释变量对因变量的影响程度。

几个难题请大家帮忙解决：

1.DW值超过2，但由于引入了变量B2？（-1），是否从DW值仍可判断自相关？

2.三个关键解释变量均有统计学意义，但是B3的符号与经济学意义相反，怎么调整呢？

请大家帮帮忙，谢谢！

………………………………………………………………………………………………

Dependent Variable: DLOG(AGRIAREA?)
Method: Pooled EGLS (Cross-section weights)
Sample (adjusted): 2000 2002
Included observations: 3 after adjustments
Cross-sections included: 31
Total pool (balanced) observations: 93
Linear estimation after one-step weighting matrix

Variable Coefficient Std. Error t-Statistic Prob.

C -0.016610 0.002280 -7.286049 0.0000
DLOG(A1?) 0.007214 0.004227 1.706637 0.0937
DLOG(A2?) -0.125493 0.015048 -8.339441 0.0000
DLOG(A3?) 0.115760 0.021278 5.440387 0.0000
DLOG(A4?) -0.043815 0.033614 -1.303474 0.1980
DLOG(A5?) 0.055073 0.014327 3.844041 0.0003
DLOG(A6?) 0.008969 0.007990 1.122588 0 .2667
DLOG(B1?) -0.034674 0.012510 -2.771690 0.0077
DLOG(B2?(-1)) -0.008383 0.001076 -7.788456 0.0000
DLOG(B3?) -0.003419 0.001701 -2.010312 0.0495
Fixed Effects (Cross)
……
Effects Specification

Cross-section fixed (dummy variables)

Weighted Statistics

R-squared 0.923611 Mean dependent var -0.033125
Adjusted R-squared 0.867400 S.D. dependent var 0.058017
S.E. of regression 0.020690 Sum squared resid 0.022687
F-statistic 16.43117 Durbin-Watson stat 2.710867
Prob(F-statistic) 0.000000

Unweighted Statistics

R-squared 0.902062 Mean dependent var -0.011384
Sum squared resid 0.029087 Durbin-Watson stat 3.044500

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

filippofox

2007-5-8 00:35:00

我还拿不准是选择dlog(x1)这样的变量呢，还是选择d(x1)这样的呢。

下面是我用后者做的一个结果：

其中的DW值更大，但是残差平方和更小，我究竟应该选择哪个呢？谢谢大家了！

Dependent Variable: D(AGRIAREA?)
Method: Pooled EGLS (Cross-section weights)
Sample (adjusted): 2000 2002
Included observations: 3 after adjustments
Cross-sections included: 31
Total pool (balanced) observations: 93
Linear estimation after one-step weighting matrix

Variable Coefficient Std. Error t-Statistic Prob.

C -0.003780 0.000765 -4.938464 0.0000
D(CAPITAL?) -9.08E-09 8.26E-09 -1.098126 0.2771
D(AGSHR?) -0.101569 0.030513 -3.328686 0.0016
D(EMPSHR?) 2.29E-05 6.98E-06 3.278122 0.0018
D(INC?) 3.26E-06 3.09E-06 1.055947 0.2958
D(DMC?) 3.47E-06 1.31E-06 2.649524 0.0106
D(HHAREA?) 0.000851 0.001215 0.700075 0.4869
D(RI?) -0.000571 0.000184 -3.107234 0.0030
D(NPR?(-1)) 0.003887 0.008534 0.455450 0.6506
D(PCOMP?) 7.37E-09 4.06E-09 1.816070 0.0750

Effects Specification

Cross-section fixed (dummy variables)

Weighted Statistics

R-squared 0.830201 Mean dependent var -0.009084
Adjusted R-squared 0.705255 S.D. dependent var 0.024993
S.E. of regression 0.004926 Sum squared resid 0.001286
F-statistic 6.644473 Durbin-Watson stat 3.060664
Prob(F-statistic) 0.000000

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

filippofox

2007-5-8 00:49:00

下边是我用Dlog（x1）这样形式的变量估计后的残差相关矩阵，这样的结果是否可以接受呢？

谢谢！

_BJ _TJ _HEBEI _SX _NM _LN _JL _HLJ

_BJ 1.000000 -0.653589 -0.999695 -0.764361 -0.956987 0.040621 -0.100109 0.071734
_TJ -0.653589 1.000000 0.672092 0.987586 0.845062 0.729675 0.818477 0.708015
_HEBEI -0.999695 0.672092 1.000000 0.780061 0.963864 -0.015918 0.124665 -0.047065
_SX -0.764361 0.987586 0.780061 1.000000 0.918557 0.613207 0.718069 0.588297
_NM -0.956987 0.845062 0.963864 0.918557 1.000000 0.251018 0.384477 0.220735
_LN 0.040621 0.729675 -0.015918 0.613207 0.251018 1.000000 0.990089 0.999514
_JL -0.100109 0.818477 0.124665 0.718069 0.384477 0.990089 1.000000 0.985232
_HLJ 0.071734 0.708015 -0.047065 0.588297 0.220735 0.999514 0.985232 1.000000

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

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