我在用论坛里的LS程序做结构突变点问题时,得到Gauss运行的结果如下:
---------------------------------------------------
Lee and Strazicich Min LM t-stat with two breaks
Model = 1
----------------------------------------------------
Model (1=A, 2=C) = 1.0000
Min. test statistic = -2.2070
Estimated break point = 21.0000 23.0000
Selected lag = 6.0000
Est. coeff. of dummy = 113.4170 1427.4394
Its t-stat = 0.3873 3.1296
Standard error .. = 256.5992
Standardized dummy = 0.4420 5.5629
Coeff and t-stat
Z(t) = [S(t-1), (lags..omitted), 1, B1(t), B2(t)]
-0.3409 -2.2070
494.8956 2.0515
113.4170 0.3873
1427.4394 3.1296
---------------------------------------------------
Lee and Strazicich Min LM t-stat with two breaks
Model = 2.0000
----------------------------------------------------
Model (1=A, 2=C) = 2.0000
Min. test statistic = -160.8174
Estimated break point = 17.0000 24.0000
Selected lag = 6.0000
Est. coeff. of dummy = 4.6263 -4004.9646
Its t-stat = 0.3676 -123.7073
Standard error .. = 5.2033
Standardized dummy = 0.8891 -769.7010
Coeff and t-stat
Z(t) = [S(t-1), (lags..omitted), 1, B1(t), B2(t), D1(t), D2(t)]
-2.0084 -160.8174
1998.2169 159.4131
-4.3877 -0.3225
4.9683 0.5031
4.6263 0.3676
-4004.9646 -123.7073
请问Model 1和2的区别是不是2的结构突变点允许同时出现在截距和斜率上?一般情况是不是以model 2的结果作为结论?
Min. test statistic 的值过大,是什么问题?怎样解释?