The Heteroskedasticity Test White with and without including cross terms give contradictory results. By including the cross terms the following results lead to rejection of null hypothesis.
Heteroskedasticity Test: White
F-statistic 2.894027 Prob. F(5,3729) 0.0130
Obs*R-squared 14.43739 Prob. Chi-Square(5) 0.0131
Scaled explained SS 1508.234 Prob. Chi-Square(5) 0.0000
On the other hand if I am not including cross terms, the results accepts the null hypotheis
Heteroskedasticity Test: White
F-statistic 2.224166 Prob. F(2,3732) 0.1083
Obs*R-squared 4.446607 Prob. Chi-Square(2) 0.1083
Scaled explained SS 464.5247 Prob. Chi-Square(2) 0.0000
Which one should I use? If I am using include cross terms and assume that heteroskedasticity is present, how shall I proceed with my multiple regression model.
Expecting clarifications. Thank you
Heteroskedasticity Test White
Moderators: EViews Gareth, EViews Moderator
-
- Non-normality and collinearity are NOT problems!
- Posts: 3775
- Joined: Wed Sep 17, 2008 2:25 pm
Re: Heteroskedasticity Test White
Include the cross-terms and use robust standard errors.
Re: Heteroskedasticity Test White
Thank You Startz
Re: Heteroskedasticity Test White
Can any one please help me in this dilemma. I have 13 variables. Eleven of them are used as dependent variables and two are common independent variables, which use independently to run 11 regression equations. However when i consider group unit root of this 13 cross sections with 41924 observation I get the following output.
Group unit root test: Summary
Series: ER_BODAL, ER_COSMO, ER_DALMBHART, ER_EXCELCO,
ER_GRAUER, ER_JINDASTEEL, ER_NCL, ER_ORI_CARB,
ER_POLYPLEX, ER_RASHTRIYA, ER_STGOBAIN,
ER_BASMAT, ER_SENSEX
Date: 02/21/17 Time: 09:34
Sample: 4/03/2001 3/31/2016
Exogenous variables: Individual effects
Automatic selection of maximum lags
Automatic lag length selection based on SIC: 0 to 1
Newey-West automatic bandwidth selection and Bartlett kernel
Cross-
Method Statistic Prob.** sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* 63.8800 1.0000 13 41924
Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat -121.281 0.0000 13 41924
ADF - Fisher Chi-square 1674.34 0.0000 13 41924
PP - Fisher Chi-square 592.280 0.0000 13 41925
** Probabilities for Fisher tests are computed using an asymptotic Chi-square
distribution. All other tests assume asymptotic normality.
Since null hypothesis of unit root cannot be rejected since p value of Levin, Lin & Chu t* statistics is >0.05, and the series are stationary using individual unit process assumption, the following doubts arise
1. Can i ignore the presence of unit root under assumption of common unit root since I am not using panel data and use regression individually.
2. If I ignore that will it in any way affect my regression results
Looking forward for getting help from anyone here
Regards
Suresh
Group unit root test: Summary
Series: ER_BODAL, ER_COSMO, ER_DALMBHART, ER_EXCELCO,
ER_GRAUER, ER_JINDASTEEL, ER_NCL, ER_ORI_CARB,
ER_POLYPLEX, ER_RASHTRIYA, ER_STGOBAIN,
ER_BASMAT, ER_SENSEX
Date: 02/21/17 Time: 09:34
Sample: 4/03/2001 3/31/2016
Exogenous variables: Individual effects
Automatic selection of maximum lags
Automatic lag length selection based on SIC: 0 to 1
Newey-West automatic bandwidth selection and Bartlett kernel
Cross-
Method Statistic Prob.** sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* 63.8800 1.0000 13 41924
Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat -121.281 0.0000 13 41924
ADF - Fisher Chi-square 1674.34 0.0000 13 41924
PP - Fisher Chi-square 592.280 0.0000 13 41925
** Probabilities for Fisher tests are computed using an asymptotic Chi-square
distribution. All other tests assume asymptotic normality.
Since null hypothesis of unit root cannot be rejected since p value of Levin, Lin & Chu t* statistics is >0.05, and the series are stationary using individual unit process assumption, the following doubts arise
1. Can i ignore the presence of unit root under assumption of common unit root since I am not using panel data and use regression individually.
2. If I ignore that will it in any way affect my regression results
Looking forward for getting help from anyone here
Regards
Suresh
Return to “Econometric Discussions”
Who is online
Users browsing this forum: No registered users and 60 guests