黃河泉 发表于 2018-4-25 15:49
没错,但为什么(你知道吗)?
具体不是太清楚,我改了控制变量从新做,结果类似,存在单门槛-----------------------------------------------------
model | Threshold Lower Upper
-----------+-----------------------------------------
Th-1 | 0.7700 0.7650 0.7710
-----------------------------------------------------
Threshold effect test (bootstrap = 300):
-------------------------------------------------------------------------------
Threshold | RSS MSE Fstat Prob Crit10 Crit5 Crit1
-----------+-------------------------------------------------------------------
Single | 4.0042 0.0131 13.02 0.0633 10.6330 13.8724 18.5789
-------------------------------------------------------------------------------
Fixed-effects (within) regression Number of obs = 315
Group variable: city Number of groups = 35
R-sq: within = 0.6664 Obs per group: min = 9
between = 0.6003 avg = 9.0
overall = 0.5957 max = 9
F(7,273) = 77.92
corr(u_i, Xb) = -0.5867 Prob > F = 0.0000
------------------------------------------------------------------------------
lhp | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lde | .628857 .1529641 4.11 0.000 .3277178 .9299962
ljgmj | .0436133 .0238796 1.83 0.069 -.0033983 .0906249
exp | .2476642 .0610651 4.06 0.000 .1274458 .3678826
inv | .3248314 .1768167 1.84 0.067 -.0232661 .6729289
lpgdp | .506828 .0371485 13.64 0.000 .433694 .579962
|
_cat#c.eff |
0 | .3708145 .2067586 1.79 0.074 -.0362295 .7778584
1 | .2093084 .1797584 1.16 0.245 -.1445805 .5631973
|
_cons | -1.774496 .9025678 -1.97 0.050 -3.551374 .0023816
-------------+----------------------------------------------------------------
sigma_u | .34174522
sigma_e | .12118656
rho | .8882976 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(34, 273) = 33.22 Prob > F = 0.0000
这个结果能够证明,当eff<0.77时,eff对lhp有正向促进作用,当>0.77时影响不显著