(EViews10):Estimate VAR Models(1)

[@CrunchEconometrix]

KZbin recently changed the way my content will be monetised. My channel now needs 1,000 subscribers. So it would be amazing if you show your support by both watching my videos and subscribing to my channel if you haven’t done so already. Monetising my videos allows me to invest back into the channel with some new equipment so this small gesture from you will be extremely huge for me. Many thanks for your support….CrunchEconometrix loves to teach, help me stay online.

[@alizaidi9]

your style of teaching is very easy to understand

[@CrunchEconometrix]

Thanks for the encouraging feedback, Ali. Deeply appreciated! May God bless you, amen!

[@lewischimfwembe3213]

I do find your Tutorials on Econometric very useful. They really assisted me during my masters degree dissertation where i used the ARDL Model. Keep doing the good work.

[@CrunchEconometrix]

Hi Lewis, thanks for the encouraging feedback. Deeply appreciated! Please may I know from where (location) you are reaching me?

[@Deskendo4341]

Meu canal mais favorito do KZbin. Obrigado por nos transmitir muito conhecimento, eu falo directamente de Moçambique e gostaria com toda gentileza a Doutora que fizesse um vídeo a explicar sobre *Teste de validação do modelo*

[@CrunchEconometrix]

Obrigado!🥰

[@strfrieue]

you know? your content very help me for graduating from college. thank you so much

[@CrunchEconometrix]

Thanks for this positive feedback, TerseSat!!!! Glad to be of great help😊

[@jinliangshan8226]

me too! Ma'am you help me get 88 in my modeling lectures and my dissertation, thank you very much.

[@omerfarukozyalcin]

Hello Sir; If we have got a dummy variable, should we put it in the part of exogenous variables part? Thank you very much for your help. Your videos are helping me so much.

[@CrunchEconometrix]

Yes Omer, ideally dummy variables should be treated as exogenous variables...and thanks for the encouraging feedback. Deeply appreciated! 🙏 ❤️

[@otzi1]

Sir değil kendisi kadın :)

[@LUCIASANTANAOLIVA]

If I want to estimate a VAR model, and I compute unit root test my variables would be I(1), hence if I am right I need to compute differences on my variables. But in the video says that VAR models estimated with variables in differences would be mis-specified. Therefore, my question is do I compute and insert the variables in differences if they are I(1) or it does not matter that they are I(1) and I introduce them into the model in levels (although they are integrated of order 1)?

[@CrunchEconometrix]

You need to watch my foundation video on VAR estimation to understand the two School of Thoughts on how to specify VAR models. Kindly check my playlist.

[@noname-re8yi]

Ma'am it will be very helpful if you demonstrate how to test multicollinearity in eviews

[@CrunchEconometrix]

Hi Nikhil, I have noted your suggestion. I will work on it. Thanks!

[@adamroble7106]

Is it necessary that the variables in the VAR system are all I(1)? I have five variable of which 2 are I(0)--stationary at levels, can i still run a VAR model?

[@CrunchEconometrix]

Hi Adam, YES for the 1st part and NO for the 2nd.

[@MiksIndrasis70]

Thank you for your work. I am very grateful to you for your help. Please help sort it out. How is a model with an optimal lag length 1 different from a model with an optimal lag length 2? What can be said about the model if its optimal lag length is 1 ? Can we make vecm model if variables are cointegrated, but optimal lag length for var is 1?

[@CrunchEconometrix]

Thanks for the positive feedback and remarks on my KZbin videos. Deeply appreciated! You will obtain different results from using different optimal lags. Yes, perform VECM if there's cointegration and you can use 1 lag if that's the optimal. Thanks for watching and sharing my videos, grateful! ❤️

[@joshuaconstantine269]

Thank you soo much for this

[@CrunchEconometrix]

You're so welcome, Joshua!

[@abhayakumar5217]

Hello Madam, Is there any video on TVP VAR. Please help us on TVP VAR. In Eviews or Stata.

[@CrunchEconometrix]

Hi Abhaya, pls what is TVP?

[@sibykm]

Madam some videos show that first differenced variables are taken as endogenour variables like d(lnpce) d(lnpdi) d(lngdp). are lnpce,lnpdi and lngdp stationary at level? or stationary at first difference?

[@CrunchEconometrix]

Hi Siby, I explained these in the video. Kindly watch again. Thanks.

[@sibykm]

@@CrunchEconometrix Thank You Madam. What I understood was that I don't need to use first differenced variables in endogenous variable box, even though they are I(1). It will be done by Eviews automatically. Please correct me if I am wrong

[@kujtimhameli]

@@sibykm I think we still need to create new differenced variables and work with them!

[@zahidahmed7837]

Mam what to do with exogeneous variables....if used only endgeneous variables...

[@CrunchEconometrix]

Hi Zahid, I'm aware that VAR also allows exogenous variables. But such will not be used as a depvar.

[@dinobrown5956]

Thanks very much. Please which estimation technique do you use to estimate a heterogeneous Panel VAR with fixed effect?

[@CrunchEconometrix]

No idea. Never done it before.

[@birendranarayanshah]

Good Morning Professor! Since variables are stationary at the first level I(1) how can we run VAR at level I(0)??

[@CrunchEconometrix]

1(0) implies stationary at level not to be confused with the level (raw) form of the series.

[@henokhaile7894]

It's an impressive tutorial by Ngozi Adeleye....What do we do if all max stastistics are less than 5% critical value ..in johansen co integration test? trace statistics result are nice

[@CrunchEconometrix]

Hi Henok, kindly watch the video on Johansen Cointegration. It answers your query. Thanks.

[@markwang2740]

Hi, if I would like to create a model of GDP, unemployment, and inflation, how do I do that on Eviews? Thanks!

[@CrunchEconometrix]

Hi Mark, depending on the technique you want to use, watch any of my EViews videos and construct your model.

[@markwang2740]

CrunchEconometrix Hi thank you! How about if I were to use it on Stata? Is it the same process as you described? I am trying to create an economic forecast model

[@CrunchEconometrix]

I have not attempted forecasting in panel data. You may need to check other online resources. Thanks.

[@sram1993]

Hello, can you please explain more why should we specified the VAR model in levels and not in differences? cause I've seen a lot people who put the variables in differences! Thank you very much for your helpful videos.

[@CrunchEconometrix]

Hi Brahim, there's a lot of discourse about that. You can look up the reference indicated. Those who estimate using the difference of the series aren't wrong going by their school of thought. Please may I know from where (location) you are reaching me?

[@sram1993]

@@CrunchEconometrix Thank so much for your response. I'm from Morocco. Please I have another question if you don't mind, I have a Var(1) model and I want to do a Granger causality tests with 2 lags it is possible to do that ?

[@CrunchEconometrix]

Alright Brahim. Estimate the VAR using 2 lags then perform the Granger causal test. I'll appreciate if you can share the link to my KZbin Channel with your colleagues in Morocco...thanks!

[@serman5671]

Hi. The link to the ex21-1.wf1 datase is broken. It would be realy helpfull if you update it. Thank you.

[@CrunchEconometrix]

Hi Ser, due to abuse and unethical conduct, datasets are accessed directly from my website since 2019. Some are free while some requires payment. Here's the link cruncheconometrix.com.ng/shop. Thank you.

[@adamuabdullahi7444]

Good day Dr. Ngozi. Thank you for another great video. I estimated a VAR at lag 1 as suggested by the lag selection criteria, and it turned out to be heteroschedastic, increase one lag made the model to be homoschedastic but I have gone against the appropriate lag length criteria, is this going to be a problem please? Lastly a monthly data of 5 years shows that the data are cointegrated through Johansen test, is best I run a VECM or VAR given the that the data although monthly is just for 5 years and what if the interest of my research is for the short run effect and the not the long run? Thank you.

[@CrunchEconometrix]

Hi Adamu, NO to the 1st question just mention what you did in your work. You have 60 observations which is ok. With cointegration estimate VECM. My videos will guide you on how to proceed. Thanks.

[@adamuabdullahi7444]

@@CrunchEconometrix thank you very much. I appreciate the guidance.

[@madelbalane369]

Hi! May I know where is the part 2 of this video?

[@CrunchEconometrix]

Hi Madel, kindly check through the EViews Time Series Playlist. Thanks.

[@aunahabdulghani3857]

Hello Prof. Currently I am working on my research topic: effect of fdi on economic sub-sectors. My dependent variable is real gdp and independent variables are fdi(major), oil price, trade openness and interbank rate. My concern is I am confused about endogenous and exogenous variables. Should i put endogenous variables in estimate VAR part: gdp and fdi only? or all of the other variables? Thank you so much.

[@CrunchEconometrix]

Hi Aunah, kindly follow the steps as shown. You may also need to read about the differences between endo and exo variables from any econometrics textbook for proper understanding. Thanks

[@enitanwale-odunaiya7350]

what do you do ma, when variables are integrated of I(1) and I(2)

[@CrunchEconometrix]

Apply the Toda-Yamamoto technique.

[@risyuwonoyudonugroho2606]

very helpfull

[@CrunchEconometrix]

Glad to hear that, Ris!

[@pedromrfernandes]

can I estimate a VAR to analyse short-run relationship with I(2) and I(1) variables? thank you very much.

[@CrunchEconometrix]

Change the I(2) variable.

[@MiksIndrasis70]

What does it mean to change in this case?

[@CrunchEconometrix]

@@MiksIndrasis70 I mean, drop the I(2) variable and use a closer proxy which can be stationary at 1st difference.

[@mishalkhaled8327]

can we use first differences instead of the level if the time series is stationary at first level only?

[@CrunchEconometrix]

If the series is stationary at level, there's nothing to test again.

[@mishalkhaled8327]

@@CrunchEconometrix they are not stationary at level. They are stationary at first level..please read my previous comment again.

[@CrunchEconometrix]

Jasem, forgive me if I have no idea of what "stationary at first level" implies. You may need to check other online resources for constructive feedback. Thanks.

[@mishalkhaled8327]

@@CrunchEconometrix at a level no stationary took the first difference became stationary so we use the first difference. thank you please be patient with your follower's comments.

[@CrunchEconometrix]

Alright, Jasem. Thanks for the clarification. The mix-up came from you. No one is perfect, though. Grateful!

[@dr.surajitdas2792]

Dr. I have a doubt... u told that data should b used in 1st diff for var nxt slide u told at level... could u plz clear the concept

[@CrunchEconometrix]

Hi there, if you watched my foundational VAR video I mentioned the 2 opinions on VAR specifications. I follow that of level specification.

[@jejedc4965]

Wait, im confused. If the variables should be integrated of order one, then why are the variables specified in levels? Shouldn't it be differenced and then specified? I've read a lot of papers about this and most of them difference the data then specify the var model.

[@CrunchEconometrix]

Hi Jeje, I gave an explanation to this regarding the school of thought on VAR specification. I suggest you watch that clip and go through the relevant literature.

[@jejedc4965]

@@CrunchEconometrix From what video is it and where can I find it? Would you mind giving the title to me? Thank you so much

[@CrunchEconometrix]

Jeje, it's the video on VAR specification. Kindly search through my Channel. Thanks.

[@saifulzan3537]

So, if my variables is in I(1) , var in level will use?

[@CrunchEconometrix]

Saiful, my explanations are clear. Watch again and follow the guide or you may seek other online tutorials.

[@saifulzan3537]

@@CrunchEconometrix thank you very much mam

[@devyaninitturkar8662]

Mam can you please tell me one thing Suppose I have dataset of different medicines(120) with their monthwise sale for 3 years then I need to use VAR method ?

[@CrunchEconometrix]

You have a panel data not a time-series data.

[@devyaninitturkar8662]

@@CrunchEconometrix mam then which statistical tood I need to apply

[@CrunchEconometrix]

Watch my panel data videos.

[@devyaninitturkar8662]

@@CrunchEconometrix ok mam thank you so much

[@syedshafiahmed8927]

Can i get ppt of the above topics?

[@CrunchEconometrix]

Not available, Syed.

[@syedshafiahmed8927]

@@CrunchEconometrix ok Ma'am Can you tell me any book where i can get step by step instructions and interpretation of ARCH/GARCH, VAR and GMM

[@CrunchEconometrix]

Hi Syed, you may not find all in one textbook.

[@devyaninitturkar8662]

Mam which software it is ?

[@CrunchEconometrix]

Hi Devyani, EViews10. Always indicated on the 1st line of the video description.

[@devyaninitturkar8662]

@@CrunchEconometrix thank you mam

[@md.mahfujurrahman864]

wao..your style is much impressive.. I am gonna fan of you...I wish I could go out with you just for a cup of coffee

[@CrunchEconometrix]

mahfujur rahman hahahaha, thanks Mahfujur! Good to hear that my videos have some level of impact...I'll appreciate if you tell others too by sharing my videos and link🥂👍🏽

[@TheSpirit99]

It would be more helpful if VAR models would be in order in the list

[@CrunchEconometrix]

How do you mean? Clarify your query.

[@homoperse979]

madam you are saying var must be specified in levels but what if all my variables become stationary after the first difference? do I need to take first difference or I must regress them on level?

[@CrunchEconometrix]

Watch my underlying video on VAR specification for better understanding of what I did.

[@santatrarakotoarimino5316]

Hey, did u find an answer to ur question ?

[@CrunchEconometrix]

Watch my videos on VAR specification and estimation. Detailed to guide you.