在使用xthreg做非平衡面板回归时,d为核心解释变量,GDP为门槛变量,跑出来的结果出现两个问题:
1、Threshold effect test中部分指标缺失,而且有前辈说RSS不应该为0,但是如果将门槛变量也作为控制变量放入时,就不会出现这种情况了
2、回归结果中,门槛只给了02,双门槛的结果汇报应该是012吧
刚接触门槛效应,在过程中遇到特别多困难,好不容易安装好了,结果运行结果又不对了,还请各位前辈指点!
. xthreg PL 市净率PB roeb lev Asset z largestholderrate age 营业收入增长率B 固定资产比率 i.year, rx(d) qx(GDP) thnum(2) trim(0.05) bs(50)
Estimating threshold within range ( 1889.5300 , 4.03e+04 ), obs = 222
1st ... 2nd ... Done
Bootstrap for single threshold
.................................................. + 50
Bootstrap for double threshold model:
.................................................. + 50
Threshold estimator (level = 95):
-----------------------------------------------------
model | Threshold Lower Upper
-----------+-----------------------------------------
Th-1 | 1889.5300 1889.5300 1945.6200
Th-21 | 2548.7100 2548.7100 2640.3101
Th-22 | 2548.7100 2548.7100 2640.3101
-----------------------------------------------------
Threshold effect test (bootstrap = 50):
--------------------------------------------------------------------------------------
Threshold | RSS MSE Fstat Prob Crit10 Crit5 Crit1
-----------+--------------------------------------------------------------------------
Single | 0.0000 0.0000 . 0.000 . . .
Double | 0.0000 0.0000 . 0.000 . . .
--------------------------------------------------------------------------------------
Fixed-effects (within) regression Number of obs = 18,286
Group variable: stkcd Number of groups = 3,883
R-squared: Obs per group:
Within = 0.0089 min = 1
Between = 0.0000 avg = 4.7
Overall = 0.0042 max = 8
F(18,14385) = 7.17
corr(u_i, Xb) = -0.2953 Prob > F = 0.0000
-----------------------------------------------------------------------------------
PL | Coefficient Std. err. t P>|t| [95% conf. interval]
------------------+----------------------------------------------------------------
市净率PB | -.1064232 .0533553 -1.99 0.046 -.2110065 -.0018399
roeb | -4.105397 1.829539 -2.24 0.025 -7.69153 -.519265
lev | 4.419401 1.206257 3.66 0.000 2.054981 6.783821
Asset | -1.533993 .3340938 -4.59 0.000 -2.18886 -.8791259
z | .5020465 .3660697 1.37 0.170 -.2154974 1.21959
largestholderrate | -.0088644 .0253623 -0.35 0.727 -.0585778 .040849
age | -.4331326 .408022 -1.06 0.288 -1.232908 .3666431
营业收入增长率B | 1.02824 .273962 3.75 0.000 .4912395 1.565241
固定资产比率 | -1.512947 1.627897 -0.93 0.353 -4.703835 1.67794
|
year |
2016 | -.1205943 .5434572 -0.22 0.824 -1.18584 .9446518
2017 | .3478441 .8997713 0.39 0.699 -1.415824 2.111512
2018 | .6930877 1.28378 0.54 0.589 -1.823286 3.209462
2019 | 1.159126 1.670663 0.69 0.488 -2.11559 4.433841
2020 | .8445936 2.066592 0.41 0.683 -3.206193 4.89538
2021 | .9424171 2.465528 0.38 0.702 -3.890335 5.77517
2022 | 1.315313 2.863894 0.46 0.646 -4.298288 6.928915
|
_cat#c.d |
0 | .3175514 .1128847 2.81 0.005 .0962829 .53882
2 | .086313 .0611544 1.41 0.158 -.0335575 .2061834
|
_cons | 23.92575 6.763967 3.54 0.000 10.66751 37.184
------------------+----------------------------------------------------------------
sigma_u | 8.6972313
sigma_e | 9.7999849
rho | .44059387 (fraction of variance due to u_i)
-----------------------------------------------------------------------------------
F test that all u_i=0: F(3882, 14385) = 2.92 Prob > F = 0.0000