|
问题1:
我在应用面板门限模型的时候得出以下结果,样本量:n=29,t=8,nt=232,从这个门槛效果自抽样中是不是可以确定存在双重门槛?
-------------------------------------------------------------------------------
临界值
------------------------------------------------------------------
模型 F值 P值 BS次数 1% 5% 10%
-------------------------------------------------------------------------------
单一门槛 18.481* 0.065 200 27.323 20.384 16.765
双重门槛 9.074*** 0.010 200 8.892 1.184 -1.851
三重门槛 0.756 0.555 200 18.133 8.495 5.656
-------------------------------------------------------------------------------
在但在结果中出现了omitted,这是什么意思,是说系数约等于0么?
---搜索第二个门槛值---
第二个门槛估计值:.00341076
---重新搜索第一个门槛值---
更新后的第一个门槛估计值:.00622386
Fixed-effects (within) regression Number of obs = 232
Group variable: idnew Number of groups = 29
R-sq: within = 0.7519 Obs per group: min = 8
between = 0.1001 avg = 8.0
overall = 0.0093 max = 8
F(6,197) = 99.51
corr(u_i, Xb) = -0.9634 Prob > F = 0.0000
------------------------------------------------------------------------------
lnt | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lnsr | -.3747769 .091804 -4.08 0.000 -.5558216 -.1937323
lnso | -.1238704 .0268825 -4.61 0.000 -.1768848 -.070856
lnsf | .048129 .0337872 1.42 0.156 -.0185021 .1147601
lntr | -.2902122 .0574944 -5.05 0.000 -.4035956 -.1768287
lnso_1 | .0989882 .0220085 4.50 0.000 .0555857 .1423906
lnso_2 | .0445079 .0114802 3.88 0.000 .021868 .0671478
lnso_3 | (omitted)
_cons | 3.886746 .4448359 8.74 0.000 3.009495 4.763998
-------------+----------------------------------------------------------------
sigma_u | 1.1301851
sigma_e | .17159799
rho | .97746658 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(28, 197) = 19.12 Prob > F = 0.0000
问题2:第二个区间出现了共线性,对每个区间进行单独估计之后第二个区间的估计系数与第一个区间的估计系数完全一致,这种情况是为什么呢?
如果是这样,还能确定是有两个门槛么?
具体的估计结果如下:
note: g2of omitted because of collinearity
------------------------------------------------------------------------------
lnt | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lnsr | -.3747769 .091804 -4.08 0.000 -.5558216 -.1937323
lnso | -.0793625 .0247266 -3.21 0.002 -.1281252 -.0305997
lnsf | .048129 .0337872 1.42 0.156 -.0185021 .1147601
lntr | -.2902122 .0574944 -5.05 0.000 -.4035956 -.1768287
g1of | .0544802 .0192834 2.83 0.005 .0164519 .0925085
g2of | (omitted)
g3of | -.0445079 .0114802 -3.88 0.000 -.0671478 -.021868
_cons | 3.886746 .4448359 8.74 0.000 3.009495 4.763998
-------------+----------------------------------------------------------------
sigma_u | 1.1301851
sigma_e | .17159799
rho | .97746658 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(28, 197) = 19.12 Prob > F = 0.0000
|