| |
TAR
|
MTAR
|
|---|
|
Model-1
|
Model-2
|
Model-3
|
Model-4
|
Model-1
|
Model-2
|
Model-3
|
Model-4
|
|---|
|
ρ+
|
−0.122 [0.037]
|
− 0.087 N [0.029]
|
− 0.104 [0.032]
|
−0.110 [0.033]
|
− 0.274 [0.055]
|
−0.172 [0.038]
|
− 0.255 [0.052]
|
−0.246 [0.048]
|
|
ρ−
|
−0.096 [0.0526]
|
−0.026 [0.038]
|
− 0.058 [0.047]
|
−0.055 [0.046]
|
− 0.051 [0.035]
|
−0.005 [0.029]
|
− 0.037 [0.029]
|
−0.026 [0.031]
|
|
\( {\hat{F}}_{Equal} \)
|
0.181 (1.374)
|
1.618a (1.364)
|
0.668 (1.438)
|
0.972 (1.358)
|
12.214a (6.704)
|
12.322a (6.900)
|
13.355a (6.786)
|
15.188a (6.799)
|
|
\( {\hat{F}}_{Joint} \)
|
6.968a (6.229)
|
4.655 (6.175)
|
5.936 (6.257)
|
6.104 (6.233)
|
13.277a (8.345)
|
10.152a (8.451)
|
12.529a (8.315)
|
13.493a (8.385)
|
|
Long run cointegration
|
√
|
×
|
×
|
×
|
√
|
√
|
√
|
√
|
|
Asymmetry effects
|
×
|
√
|
×
|
×
|
√
|
√
|
√
|
√
|
- Note: a denote significance at 10% level. TAR indicate threshold autoregressive model with threshold indications. Numbers in [] and () are standard errors and the Enders and Siklos (2001) bootstrap critical value
