@josephlanre Sir, I appreciate your reply. I have checked that video of you and run regression based on that code. But it's not working. If I may request you, please make a detailed video on Stata or eview for the panel.
Hi Sir. May i know if these steps are applicable to the threshold model developed by Khan & Senhadji (2001)? Another question is also applicable for time series that have stationary variables?
Hello Lanre, Thanks for this video. However as a beginner, I have been trying to set up the TAR equation and I have been unsuccessful with it. Could you put up an introductory video
@@josephlanre Thank you. Do you know how to install demo version without buying it? Because I need only one model TAR for my paper and after I won't use it at all.
Hi Hanif. The slope change interns of magnitude does not necessarily mean nonlinearity. Nonlinearity comes in using this method in terms of the signs. E.g. If it is positive at certain threshold, and then negative at another, we have inverted U-shape relationship and also nonlinear relationship. It if is however positive or negative at all threshold, then, there is linearity in the model. You assertion would be more valid using Nonlinear ARDL tech.
@@josephlanre Thank you so much. Sir, how about if it is negative and insignificant at a certain threshold, and then negative and significant at another?
@@josephlanre Thank you so much sir. Are there any previous research work which states that stationary isn’t required. That is in order to mention in my research paper.
@@pujaneeparanayapa3461 you can look at the work of Tjostheim and Yin(2011) on estimation in threshold autoregressive models with a stationary and a unit root regime
@@josephlanre Thank you sir. I’ll refer that research paper. Sir, Are there any test to be carried out before running the TAR model. Example :in Koffi Pokou (2020) mentions about F (gamma ) test that is carried out to test the null hypothesis on whether there is a linear relationship between the variables where the alternative hypothesis is presence of threshold effect. Further, LM statistic also has been used to test the hypothesis on the absence of a threshold effect. I have searched on these test but couldn’t find any information. Kindly assist me.
@@pujaneeparanayapa3461 I'm so sorry for replying late. I was on a trip. I'm not aware of any pre-estimation test, but post estimation test might be required. The F-stat will come with the regression result. It is expected to be significant. You can check the residual diagnostic for LM test. Thank you
Good job bro.very educative.more grace
Thank you very much Joseph for a very clear explanation.Cheers. Phil.
Hello Sir, Could you please make another video or explain, how can we estimate threshold analysis in EViews or Stata for panel data?
Hi. Please, check my previous videos. I had uploaded a video on panel threshold on stata
@josephlanre Sir, I appreciate your reply. I have checked that video of you and run regression based on that code. But it's not working. If I may request you, please make a detailed video on Stata or eview for the panel.
Hi fatimoh. Thanks for the comment. You can reach me directly via this mail address. joecee001@gmail.com
Hi Sir. May i know if these steps are applicable to the threshold model developed by Khan & Senhadji (2001)? Another question is also applicable for time series that have stationary variables?
Hello Lanre,
Thanks for this video.
However as a beginner, I have been trying to set up the TAR equation and I have been unsuccessful with it.
Could you put up an introductory video
Alright Priscilla. I ll look into it
@@josephlanre hy Lanre, how to estimate treshold on panel data?
Thanks. So, what is the threshold?
From the video tutorial, it is 17.5
Hello, could you tell please which version Eviews do you use? I have 7.2, but it doesn't have Threshold autoregressive model
Hi Renata.
I used Eviews 9 for this video
@@josephlanre Thank you. Do you know how to install demo version without buying it? Because I need only one model TAR for my paper and after I won't use it at all.
@@renatatsoi4546 go to getintopc.com download eviews version 10 and during installation write demo in serial no. place
Hi, can we change the number of thresholds in eviews to 2?
can someone upload a video that shows steps of Momentum threshold autoregressive (M-TAR) model? thanks in advance
thanks , but slope has changed interms of magnitude, doesn't that make it non linear relationship ?
Hi Hanif. The slope change interns of magnitude does not necessarily mean nonlinearity. Nonlinearity comes in using this method in terms of the signs. E.g. If it is positive at certain threshold, and then negative at another, we have inverted U-shape relationship and also nonlinear relationship. It if is however positive or negative at all threshold, then, there is linearity in the model. You assertion would be more valid using Nonlinear ARDL tech.
@@josephlanre thanks :)
@@josephlanre Thank you so much. Sir, how about if it is negative and insignificant at a certain threshold, and then negative and significant at another?
Is there a requirement that the variables need to be stationery to run TAR model
Please reply. It’ll be a huge favour
It doesn't matter whether stationary or not. TAR does not involve issues around short run and long run
@@josephlanre Thank you so much sir.
Are there any previous research work which states that stationary isn’t required. That is in order to mention in my research paper.
@@pujaneeparanayapa3461 you can look at the work of Tjostheim and Yin(2011) on estimation in threshold autoregressive models with a stationary and a unit root regime
@@josephlanre Thank you sir. I’ll refer that research paper.
Sir, Are there any test to be carried out before running the TAR model.
Example :in Koffi Pokou (2020) mentions about F (gamma ) test that is carried out to test the null hypothesis on whether there is a linear relationship between the variables where the alternative hypothesis is presence of threshold effect.
Further, LM statistic also has been used to test the hypothesis on the absence of a threshold effect.
I have searched on these test but couldn’t find any information. Kindly assist me.
@@pujaneeparanayapa3461 I'm so sorry for replying late. I was on a trip. I'm not aware of any pre-estimation test, but post estimation test might be required. The F-stat will come with the regression result. It is expected to be significant. You can check the residual diagnostic for LM test. Thank you
Could not understand your English . I think the video must be good.
Please can you furnish me your email address ,I wish to discuss some certain things and challenges encounter while analysiing Nonlinear regression
Thank you for the video.what of a situation where two regime has positive values?