黄老师您好,我最近做门槛回归模型,结果是两个门槛值,LR图呈现出这样的结果,请问正常吗?谢谢您!
这是三门槛结果:
Threshold estimator (level = 95):
-----------------------------------------------------
model | Threshold Lower Upper
-----------+-----------------------------------------
Th-1 | 0.0249 0.0246 0.0250
Th-21 | 0.0249 0.0246 0.0250
Th-22 | 0.0934 0.0915 0.0958
Th-3 | 0.0283 0.0280 0.0284
-----------------------------------------------------
Threshold effect test (bootstrap = 600 600 600):
-------------------------------------------------------------------------------
Threshold | RSS MSE Fstat Prob Crit10 Crit5 Crit1
-----------+-------------------------------------------------------------------
Single | 29.8413 0.0735 73.36 0.0000 26.0408 30.9832 46.3738
Double | 28.0203 0.0690 26.39 0.0283 19.6553 23.3659 34.7658
Triple | 26.9449 0.0664 16.20 0.3417 25.5346 30.0001 44.3577
-------------------------------------------------------------------------------
Fixed-effects (within) regression Number of obs = 420
Group variable: city Number of groups = 30
R-sq: within = 0.9054 Obs per group: min = 14
between = 0.2815 avg = 14.0
overall = 0.4112 max = 14
F(8,382) = 457.01
corr(u_i, Xb) = -0.4291 Prob > F = 0.0000
------------------------------------------------------------------------------
lnzl | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
gov2 | 1.043383 .401767 2.60 0.010 .2534311 1.833335
urban | 6.34812 .6536341 9.71 0.000 5.062949 7.633291
ins | 2.261141 .3769646 6.00 0.000 1.519955 3.002326
eco1 | .3082913 .0393113 7.84 0.000 .2309977 .385585
|
_cat#c.lnfdi |
0 | .0284263 .028904 0.98 0.326 -.0284045 .085257
1 | .0486613 .0287346 1.69 0.091 -.0078366 .1051592
2 | .0643686 .0280699 2.29 0.022 .0091776 .1195595
3 | .0393711 .0287897 1.37 0.172 -.0172351 .0959772
|
_cons | 3.388653 .461902 7.34 0.000 2.480464 4.296841
-------------+----------------------------------------------------------------
sigma_u | 1.3237552
sigma_e | .26558685
rho | .96130465 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(29, 382) = 101.50 Prob > F = 0.0000
这是双门槛
Threshold estimator (level = 95):
-----------------------------------------------------
model | Threshold Lower Upper
-----------+-----------------------------------------
Th-1 | 0.0249 0.0246 0.0250
Th-21 | 0.0249 0.0246 0.0250
Th-22 | 0.0934 0.0915 0.0958
-----------------------------------------------------
Threshold effect test (bootstrap = 600 600):
-------------------------------------------------------------------------------
Threshold | RSS MSE Fstat Prob Crit10 Crit5 Crit1
-----------+-------------------------------------------------------------------
Single | 29.8413 0.0735 73.36 0.0000 24.9826 31.6139 42.0711
Double | 28.0203 0.0690 26.39 0.0567 22.0697 26.9517 34.0600
-------------------------------------------------------------------------------
Fixed-effects (within) regression Number of obs = 420
Group variable: city Number of groups = 30
R-sq: within = 0.9016 Obs per group: min = 14
between = 0.2831 avg = 14.0
overall = 0.4050 max = 14
F(7,383) = 501.46
corr(u_i, Xb) = -0.4734 Prob > F = 0.0000
------------------------------------------------------------------------------
lnzl | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
gov2 | .9441214 .4083507 2.31 0.021 .1412316 1.747011
urban | 6.95562 .6465453 10.76 0.000 5.684397 8.226842
ins | 2.331766 .3834689 6.08 0.000 1.577798 3.085734
eco1 | .3173374 .0399662 7.94 0.000 .2387569 .395918
|
_cat#c.lnfdi |
0 | .0464588 .0290584 1.60 0.111 -.0106752 .1035927
1 | .0748389 .0284565 2.63 0.009 .0188884 .1307893
2 | .0467009 .0292578 1.60 0.111 -.0108252 .1042269
|
_cons | 2.817225 .4461777 6.31 0.000 1.939961 3.69449
-------------+----------------------------------------------------------------
sigma_u | 1.3636343
sigma_e | .27048097
rho | .96214543 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(29, 383) = 102.12 Prob > F = 0.0000