Panel Unit root

[rose]

in the LLC Unit root test , every time you do the unit root tet on the same series , the results are chaning becasues it is changing the number of Cross sections included in the estimation.


First Time


Null Hypothesis: Unit root (common unit root process)
Series: GDP
Date: 04/12/11 Time: 22:24
Sample: 1987 2008
Exogenous variables: Individual effects
Automatic selection of maximum lags
Automatic lag length selection based on SIC: 0
Newey-West automatic bandwidth selection and Bartlett kernel
Total (balanced) observations: 42
Cross-sections included: 2 (2 dropped)


Method Statistic Prob.**
Levin, Lin & Chu t* 5.82132 1.0000


** Probabilities are computed assuming asympotic normality

Intermediate results on GDP


Cross 2nd Stage Variance HAC of Max Band-
section Coefficient of Reg Dep. Lag Lag width Obs
1 0.02820 1.E-05 4.E-05 0 4 2.0 21
2 Dropped from Test
3 Dropped from Test
4 0.01802 7.E-05 8.E-05 0 4 1.0 21

Coefficient t-Stat SE Reg mu* sig* Obs
Pooled 0.02703 5.169 1.005 -0.554 0.919 42





Second Time


Null Hypothesis: Unit root (common unit root process)
Series: GDP
Date: 04/12/11 Time: 22:24
Sample: 1987 2008
Exogenous variables: Individual effects
Automatic selection of maximum lags
Automatic lag length selection based on SIC: 0
Newey-West automatic bandwidth selection and Bartlett kernel
Total (balanced) observations: 84
Cross-sections included: 4


Method Statistic Prob.**
Levin, Lin & Chu t* 5.99013 1.0000


** Probabilities are computed assuming asympotic normality

Intermediate results on GDP


Cross 2nd Stage Variance HAC of Max Band-
section Coefficient of Reg Dep. Lag Lag width Obs
1 0.02820 1.E-05 4.E-05 0 4 2.0 21
2 0.02741 0.0002 0.0001 0 4 6.0 21
3 -0.11052 0.2154 0.2163 0 4 1.0 21
4 0.01802 7.E-05 8.E-05 0 4 1.0 21

Coefficient t-Stat SE Reg mu* sig* Obs
Pooled 0.02673 5.227 1.013 -0.554 0.919 84



Third Time

Null Hypothesis: Unit root (common unit root process)
Series: GDP
Date: 04/12/11 Time: 22:24
Sample: 1987 2008
Exogenous variables: Individual effects
Automatic selection of maximum lags
Automatic lag length selection based on SIC: 0
Newey-West automatic bandwidth selection and Bartlett kernel
Total (balanced) observations: 63
Cross-sections included: 3 (1 dropped)


Method Statistic Prob.**
Levin, Lin & Chu t* 2.12907 0.9834


** Probabilities are computed assuming asympotic normality

Intermediate results on GDP


Cross 2nd Stage Variance HAC of Max Band-
section Coefficient of Reg Dep. Lag Lag width Obs
1 Dropped from Test
2 0.02741 0.0002 0.0001 0 4 6.0 21
3 -0.11052 0.2154 0.2163 0 4 1.0 21
4 0.01802 7.E-05 8.E-05 0 4 1.0 21

Coefficient t-Stat SE Reg mu* sig* Obs
Pooled 0.01938 1.548 1.013 -0.554 0.919 63

[EViews Gareth]

What's the build date of your copy of EViews (available from Help->About EViews)?

[rose]

Dec 4 2009

[EViews Gareth]

Update to the latest version