Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below: ua-cam.com/video/JbXHQNazvYU/v-deo.html
Hi there, Thank you so much for your lectures. I am working on a data set of stocks return which I have already converted into log returns in Excel then imported it in the Eviews. I am facing the problem of heteroskedasticity even after changing it the estimation option at Huber-white and then Breush pagan test which shows improved values but still heteroskedastic. I am confused with if I can use log log model or not because the data has already been transformed into log model. however if I use Harvey method my results seems to have homoskedacity. waiting for your reply. Thank you so much for the good work.
Don't make log transformation twice. Log transformation compress the range o variables, and also the variance of residuals. But it's not a flawless solution against heteroskedasticity. If it were, so there were no need for robust covariance matrix, like designed by White (1980), or generalized least squares. Take easy on log transformation, man, it's not gonna solve all forms of heteroskedasticity.
is the result still reliable/robust? as we can see, the result is different, i.e Log(EX) is not significant, meanwhile EX has negative impact and significant.
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below: ua-cam.com/video/JbXHQNazvYU/v-deo.html
what if View/Residual Diagnostics/Heteroskedasticity Tests. not showing on? I mean I have Eviews 9 and don't have heteroskedasticity menu on tab view, please help
Sir can I convert a variable into log form before doing any test like unit root test and multicollinearity? Can I convert log form for dependent variable as well?
Well, that's a ungreatful question because in a VAR model there's as many residual series as variables in the model. So, every residual series could have it's own form of heteroskedasticity. If it's hard to identifiy the form of heteroskedasticity in a single residual series, then figure out how hard it would be to do it for, let's say, 3 or 4 different residual series. I would suggest you to estimate every equation of your VAR separetely and then apply White's (1980) heteroskedasticity robust covariance matrix in every single one which in there is heteroskedasticity.
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below: ua-cam.com/video/JbXHQNazvYU/v-deo.html
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below: ua-cam.com/video/JbXHQNazvYU/v-deo.html
You can't apply log in nonipositive numbers. Try to identifiy the form of heteroskedasticity. If you're able to do that, so apply the GLS method. If you're, then apply White's (1980) heteroskedasticity robust covariance matrix.
One more question, sir can I convert one or 2 variable into log form only if I want ? Or Should i convert all of them if I want to log form ?? Sir please reply
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below: ua-cam.com/video/JbXHQNazvYU/v-deo.html
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below: ua-cam.com/video/JbXHQNazvYU/v-deo.html
Very nice video and very helpful
Very good! God bless your life
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below:
ua-cam.com/video/JbXHQNazvYU/v-deo.html
What about autocorrelation in log model
Hi there, Thank you so much for your lectures. I am working on a data set of stocks return which I have already converted into log returns in Excel then imported it in the Eviews. I am facing the problem of heteroskedasticity even after changing it the estimation option at Huber-white and then Breush pagan test which shows improved values but still heteroskedastic. I am confused with if I can use log log model or not because the data has already been transformed into log model. however if I use Harvey method my results seems to have homoskedacity. waiting for your reply. Thank you so much for the good work.
Don't make log transformation twice.
Log transformation compress the range o variables, and also the variance of residuals.
But it's not a flawless solution against heteroskedasticity. If it were, so there were no need for robust covariance matrix, like designed by White (1980), or generalized least squares.
Take easy on log transformation, man, it's not gonna solve all forms of heteroskedasticity.
If you can identify the form of heteroskedasticity, so apply GLS. If you can't, then use White's robust for heteroskedasticity covariance matrix.
is the result still reliable/robust?
as we can see, the result is different, i.e Log(EX) is not significant, meanwhile EX has negative impact and significant.
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below:
ua-cam.com/video/JbXHQNazvYU/v-deo.html
what if View/Residual Diagnostics/Heteroskedasticity Tests. not showing on? I mean I have Eviews 9 and don't have heteroskedasticity menu on tab view, please help
same problem
Sir can I convert a variable into log form before doing any test like unit root test and multicollinearity? Can I convert log form for dependent variable as well?
Yes you can
This is it, log transformation?
how d o you remove heteroskedasticy in a VAR model?
Well, that's a ungreatful question because in a VAR model there's as many residual series as variables in the model. So, every residual series could have it's own form of heteroskedasticity. If it's hard to identifiy the form of heteroskedasticity in a single residual series, then figure out how hard it would be to do it for, let's say, 3 or 4 different residual series.
I would suggest you to estimate every equation of your VAR separetely and then apply White's (1980) heteroskedasticity robust covariance matrix in every single one which in there is heteroskedasticity.
Hi Sir ,how about my model shows 'non positive number'?? what mean by that
You cannot log negative no.
@@magizkise2747 thank you so much
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below:
ua-cam.com/video/JbXHQNazvYU/v-deo.html
What to do with non positive values.?
I skipped that varibles but still the problem of hetero exists
Keep guide me
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below:
ua-cam.com/video/JbXHQNazvYU/v-deo.html
You can't apply log in nonipositive numbers.
Try to identifiy the form of heteroskedasticity. If you're able to do that, so apply the GLS method. If you're, then apply White's (1980) heteroskedasticity robust covariance matrix.
One more question, sir can I convert one or 2 variable into log form only if I want ? Or
Should i convert all of them if I want to log form ?? Sir please reply
It depends on your variables
what happen if i already log all of the variables but still have heteroskedasticity problem? how can i solve it?
same question please reply
Use robust standard errors... If you don't know i will make a video in 2 3 days
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below:
ua-cam.com/video/JbXHQNazvYU/v-deo.html
Please check the built-in method of How to detect and remove heteroskedasticity in eviews from the link given below:
ua-cam.com/video/JbXHQNazvYU/v-deo.html