Dear Statalist,
I am trying to find break point and cointegration . I applied the Geroge Hansen test , according to the link ,http://www.stata.com/statalist/archi.../msg00075.html
The commands for 3 models provided in the link above work perfectly. However ,George Hansen test consists of 4 models . Model 2 (level shift and trend) for George Hansen test is still missing
Model 2: Cointegration with a level shift and trend
Yt = α + β*Xt + γ*dumt + μ*t + εt
May I ask if any of you knows the command for model 2, could you please give me any advice?.Thank you so much!!
I am writing down the equation for each model and corresponding command in the above link for your reference
Model 1: Cointegration with a level shif
Yt = α + β*Xt + γ*dumt + εt (1)
Model 3: Cointegration with a regime shift
Yt = α + β*Xt + γ*dumt + μ*t + δ* Xt*dumt + εt (3)
Model 4: Cointegration with a regime shift and trend
Yt = α + β1*Xt + γ*dumt + μ*t + φ *dumt *t + δ* Xt*dumt + εt (4)
I am trying to find break point and cointegration . I applied the Geroge Hansen test , according to the link ,http://www.stata.com/statalist/archi.../msg00075.html
The commands for 3 models provided in the link above work perfectly. However ,George Hansen test consists of 4 models . Model 2 (level shift and trend) for George Hansen test is still missing
Model 2: Cointegration with a level shift and trend
Yt = α + β*Xt + γ*dumt + μ*t + εt
May I ask if any of you knows the command for model 2, could you please give me any advice?.Thank you so much!!
I am writing down the equation for each model and corresponding command in the above link for your reference
Model 1: Cointegration with a level shif
Yt = α + β*Xt + γ*dumt + εt (1)
Code:
ghansen y x, break(level) lagmethod(aic) maxlags(5)
Model 3: Cointegration with a regime shift
Yt = α + β*Xt + γ*dumt + μ*t + δ* Xt*dumt + εt (3)
Code:
ghansen y x, break(regime) lagmethod(fixed) maxlags(5)
Yt = α + β1*Xt + γ*dumt + μ*t + φ *dumt *t + δ* Xt*dumt + εt (4)
Code:
ghansen y x, break(regimetrend) lagmethod(downt) level(0.99) trim(0.1)
Comment