Thank you sir for such a helping video, my question is what to do next when VECM shows there is no long run causality plus no short run causality as well????? Thank you
Thank you Lokki, I would like to invite you to join Hossain Academy Facebook at below link and post your question there if you have any. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Thanks a lot. However, I wonder why we went to the trouble of running a VECM only to discover that there is no long-run association yet the Johanssen Cointegration test led us to believe that there was one. Were we misled?
massive thanks for this amazing video, my question is, if it just have one problem which is the data have serial correlation and you said that was unacceptable, so what should I do? and if there is no further step anymore, what can I conclude sir? hope you reply
+Muhammad Hafiz Thank you. I would like to invite you to join Hossain Academy Facebook for greater interaction about economics, finance and econometrics with me. Thank you Sayed Hossain from Hossain Academy. Please join below and post your question .facebook.com/groups/hossainacademy/
I followed your analysis thoroughly; it is very interesting . I got so much knowledge from you. Thank you very much! I have one Question : If there is serial correlation problem or hetroskedsatisity problem and Normality problem, then how can I CORRECT THOSE PROBLEMS?
Thank you very much for your support! I followed your video of all models of stata and all models of Eview. They are very interesting. Your pronunciation and ways of explanation really push me to further follow you. I am Developmental Economic post graduate student. Now I need to get some video from you that related to stata. Sayed, would you post in my gmail.com account about ""creating new variables in stata, generate, replace, lablel variable,label define, label values, and other related issues"" Thank you! jzk
Thank Sir for all ur valuable documents, I am from Africa and im really happy to get ur explanations here, But one thing i want to ask is all u have told is the existence of the causality only concerning the longrun and short run causality and also u tested the wald test and all that... but how can we build the formula regression/equation for both (shortrun and longrun equations ) in one of your video i have seen one formula but it does not include the constant and the sign is opposite to what u see on the vecm result.Lastly i have seen how to get probability of variables in the short run by the system u showd (vecm-proc-make sytem-order variable.....) but how about the probability of longrun eqution variables? shoud i take the opposite of the sign i get vecm in shortrun also like u did in longrun equation building? .... Anyways u are doing great job Be blessed SIR
Dear Mishra, I would like to invite you to join Hossain Academy Facebook at below link to discuss about economics, econometrics and statistical models using EVIEWS, STATA, R, SPSS, Minitab, Microfit, Lingo, and Excel. Thank you, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Thank you so much Mr Hossain, it's so helpful, could help me with a question ?, I have a VECM in wich the sign of ec term is not negative, can I affirm long run causality ?, mmm, by the way, It would be wonderfull if you can do the same with programming commands... thanks, and congratulation its an excelent channel... (y)... greetings from Ecuador
Thank you. I would like to invite you to join Hossain Academy Facebook Group at below link and join our group discussion. Thank you. Sayed Hossain from Hossain Academy.
Thank you Sanjeev, I would like to invite you to join Hossain Academy Facebook at below link and post your question there. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Dear Bee, Thank you. I would like to invite you to join Hossain Academy Facebook at below link and post your question there. Actually I am in that group and may help you. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Dear Sir, Thank you very much for your good video, it is very useful for me. I have some question when I run VECM and I hope you can use a little time to respones to me. I run the model with independent variable is Trade balance (TB) and dependent variables are industrial production of Vietnam (IPVN), industrial production of China (IPCN) and real exchange rate (RER). My data have over 140 observations and they are percentage change year over year. My question are: 1. How many observations should we have when running VECM? 2. When I run Johansen test, I have the results: Trace test indicates 3 cointegrating eqn(s) at the 0.05 level and Max-eigenvalue test indicates 1 cointegrating eqn(s) at the 0.05 level. Which test should I use to run VECM and what are the different? 3. Based on the Johansen test, can I conclude that there are the long run relationship between 3 dependent variables to TB? 4. Does Wald test use to test short run causality? Thank you very much and I am looking forward to hearing from you soon. Best regards!
Cuong Tran Quoc Khanh Please join hossain academy group in the facebook and make question. Follow the following link.@login.php?next=https%3A%2F%2Fwww.facebook.com%2Fgroups%2F841025519270747%2F843415589031740%2F%3Fcomment_id%3D843444675695498%26notif_t%3Dlike
Good day Prof. thank you so much for this teaching very interesting and explanatory. Am working on minimizing risk of precious metals. is it possible to use VAR model, VECM or which method would be most appropriate. Thanks
+Dare Jayeola VAR is not suitable for that to minimize risk. However, Thank you. I would like to invite you to join Hossain Academy Facebook for greater interaction about economics, finance and econometrics with me. Thank you Sayed Hossain from Hossain Academy. Please join below and post your question.facebook.com/groups/hossainacademy/
Dear Sir thank you very much I have a question i want to know the relationship between credit and market and i have some control variables like EPS ratio, loan growth rate, total assets, return on assets. My question is do all variables have to be non stationary to run VECM or i can ignore control variables stationarity and focus on market and credit risk only??
Heba Mohamed In VECM model all variables should be non stationary at level but after first differenced they should be stationary. Are u using EVIEWS? or other software?
E views however, i am having trouble to set my model although i completely understand your video. May i tell you my variables and model may be you can help me understand how to run it...thank you sir
I watched them twice. my problem is that not all of my variables are non stationary on level. second i got 12 banks from, china, Malaysia, South Africa, Egypt, Indonesia, and Brazil. What i want to do is to see if credit risk and market risk are affected by each other and my control variables are: GDP Growth rate CPI Return on Assets Total Assets Efficiency Ratio Loan growth rate Earnings per share I am flexible in dropping any of my control variables. Is running an unrestricted VAR as shown in previous videos suitable for me if they are not all non stationary at level? Thank you very much you are a life saver cannot be thankful enough
AOA Sir; i have 9 variables in my model including dependent variable. If C(1)is negative and probability is significant at 5% level by checking long run causality but overall probability of overall model is not significant at 5% level then what should we do for making our model significant?
Dear Sir Sayed. your lectures are very helpful for me. i must say thanks to you. here i have some issue that in my model i take 3 lags and there is existence of long-run and also short run causality. what should i do ?
Dear sir Thank you so much for such a nice delivery of content. I have a question that it C(1) is -ve and insignificant than what would be our conclusion about long run causality
Thank you. I would like to invite you to join Hossain Academy Facebook at below link and post your question there. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
on the one hand the johnson test suggest there is long run relationship among variables, on the other hand, we find no casuality both in the short run and long run. How are these two results compatible??
Thank you. I would like to invite you to join Hossain Academy Facebook Group at below link and join our group discussion. Thank you. Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
I have 7 variable which 2 of them non stationer in level but stationer in 1st difference... can I use VAR? because there is condition for using VAR which is there is no cointegration
In VECM, all variables are I(1). I got rank(2). but when comes to the VEC equation, I got many insignificant results if I keep all variables like that ( meaning that state turns into D1). when change into D.varname ( and stata turns into D2.varname) I got amazingly good results. what to do? keep varnames or change into D.varname before running VECM. thanks
Zoundi Zakaria Please join Hossain Academy Facebook below and post your question. Indeed I am there to share with you. Thank you Sayed Hossain from Hossain Academyfacebook.com/groups/hossainacademy/
Dear Sir. All your videos are crucial and easy to understand. Thanks for it My paper use Eview for running the long-run relationship betwen Exchange Rate anh Stock index with considering Macro index (CPI, unemployment). I had used all test ADF, Granger and Johenson. With all your Video i knew how to do and got the result the same what you did it. I want to thank you alot. However, i have stucked with final step, using Nonparametric Cointergrating regression. Do you know it and how to run it. May you guide me how to run if you know. Thanks Sir Jane
Le Jane Hi Jane, please join Hossain Academy facebook group where we can discuss. Thanks. See below link.facebook.com/login.php?next=https%3A%2F%2Fwww.facebook.com%2Fgroups%2F841025519270747%2F843415589031740%2F%3Fcomment_id%3D843444675695498%26notif_t%3Dlike
+Naveed Raza Not yet done. However, I would like to invite you to join Hossain Academy Facebook for greater interaction about economics, finance and econometrics with me. Thank you Sayed Hossain from Hossain Academy. Please join below and post your question.facebook.com/groups/hossainacademy/
sir, when i try to run vecm ...it shows insufficient number ofobservations.i hv data of fdi and 4 more variables annual data from 1991-2018. kindly guide
what happen if the cointegration number is different for trace test (7) and max eigenvalue (4)? which of the value should i choose to run the program? 4 or 7?
Dear Gunadi, I would like to invite you to join Hossain Academy Facebook at below link and post this question there, I shall respond. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Sayed Hossain Solved it, thank you (If someone else has the same problem: eviews wasn't opening my time series properly for some reason. A restart solved it) Another quick question if you have the time: I have 3 time series, and the Johansen test (trace test) indicates 3 cointegrating equations. However, when setting up the VECM, Eviews does not permit cointegration rank 3, only 1 or 2. Must the cointegrating rank in the VECM necessarily be lower than the number of variables? Thank you.
sorry, if the model have trend (not to chosen 3, chosen 5(intercept and trend in CE-linear trend in VAR)), the order by veraible is not working. error: illegal date :) How can i understand my veriable is signifficant?
Mr. Hossain, please i want to know when i shall use Engle-Granger or Johansen? And why do you use OLS after running VECM, it isn`t enought to get the VECM?? So basically the process is: Test unit root, cointegration test,VAR, VECM and OLS? ... Please answer me it`s important, thank you!
Thank you. I would like to invite you to join Hossain Academy Facebook Group (Data Analysis) at below link and join our group discussion about modelling. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
my all the projected variables are stationary at first difference and run johanson co integration test where the trace test and eigan value test both are indicating 4 co integration equations. when i run the vecm where i slected the 4 co integration equation but the results are not producing it is giving error that co integration eq selection is not correct. what is the problem here? kindly guide.
Thank you. I would like to invite you to join Hossain Academy Facebook Group at below link and join our group discussion. Thank you. Sayed Hossain from Hossain Academy.
I Im so match satisfied in Sayed Hossain training in how to use Eview. God bless you. Thank you very match.
highly respect to you, sir. Thanks for everything that you did for spread knowledge.
Great video and a great class. Thank you!
Thank you very much for your presentation. You have made it clear step by step. It's very helpful for analysis.
Oh. that's sound great. Thank you very much. God bless you.
Thank you sir for such a helping video, my question is what to do next when VECM shows there is no long run causality plus no short run causality as well????? Thank you
Excelent video. Thank you.
You are a good teacher Mr Hossain
Thank you Lokki, I would like to invite you to join Hossain Academy Facebook at below link and post your question there if you have any. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Dear sir thankuuuu so much for the valuable information with easy steps n explaination. Great sir🙏
Are u econometric srudent
Good job Sayed. Well done!
Thanks a lot.
However, I wonder why we went to the trouble of running a VECM only to discover that there is no long-run association yet the Johanssen Cointegration test led us to believe that there was one. Were we misled?
Thanks so much!! You saved my semester
massive thanks for this amazing video, my question is, if it just have one problem which is the data have serial correlation and you said that was unacceptable, so what should I do? and if there is no further step anymore, what can I conclude sir? hope you reply
+Muhammad Hafiz Thank you. I would like to invite you to join Hossain Academy Facebook for greater interaction about economics, finance and econometrics with me. Thank you Sayed Hossain from Hossain Academy. Please join below and post your question
.facebook.com/groups/hossainacademy/
I followed your analysis thoroughly; it is very interesting . I got so much knowledge from you. Thank you very much! I have one Question : If there is serial correlation problem or hetroskedsatisity problem and Normality problem, then how can I CORRECT THOSE PROBLEMS?
+Mohammed Hebo First convert all your variables into natural log and run your model. I guess many problem will solve
Thank you very much for your support! I followed your video of all models of stata and all models of Eview. They are very interesting. Your pronunciation and ways of explanation really push me to further follow you. I am Developmental Economic post graduate student. Now I need to get some video from you that related to stata. Sayed, would you post in my gmail.com account about ""creating new variables in stata, generate, replace, lablel variable,label define, label values, and other related issues"" Thank you! jzk
Thank Sir for all ur valuable documents, I am from Africa and im really happy to get ur explanations here,
But one thing i want to ask is all u have told is the existence of the causality only concerning the longrun and short run causality and also u tested the wald test and all that... but how can we build the formula regression/equation for both (shortrun and longrun equations ) in one of your video i have seen one formula but it does not include the constant and the sign is opposite to what u see on the vecm result.Lastly i have seen how to get probability of variables in the short run by the system u showd (vecm-proc-make sytem-order variable.....) but how about the probability of longrun eqution variables? shoud i take the opposite of the sign i get vecm in shortrun also like u did in longrun equation building? ....
Anyways u are doing great job
Be blessed SIR
Thanks a lot, but got a question, Can we apply VECM if model is not normally distributed?
Thank you so much for nicely explaining the model. I have only one doubt if in VAR the lag is "k", then in VECM the lag should be "k-1" or "k"?
Yes
Dear Mishra, I would like to invite you to join Hossain Academy Facebook at below link to discuss about economics, econometrics and statistical models using EVIEWS, STATA, R, SPSS, Minitab, Microfit, Lingo, and Excel. Thank you, Sayed Hossain from Hossain Academy.
facebook.com/groups/hossainacademy/
Thank you so much Mr Hossain, it's so helpful, could help me with a question ?, I have a VECM in wich the sign of ec term is not negative, can I affirm long run causality ?, mmm, by the way, It would be wonderfull if you can do the same with programming commands... thanks, and congratulation its an excelent channel... (y)... greetings from Ecuador
You are welcome
Thank you very much !
Thank you. I would like to invite you to join Hossain Academy Facebook Group at below link and join our group discussion. Thank you. Sayed Hossain from Hossain Academy.
NICE EXPLANATION SIR....
Thank you Sanjeev, I would like to invite you to join Hossain Academy Facebook at below link and post your question there. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Thank you so much. You are a Life saver :-)
Dear Bee, Thank you. I would like to invite you to join Hossain Academy Facebook at below link and post your
question there. Actually I am in that group and may help you. Thank you once
again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Dear Sir,
Thank you very much for your good video, it is very useful for me.
I have some question when I run VECM and I hope you can use a little time to respones to me.
I run the model with independent variable is Trade balance (TB) and dependent variables are industrial production of Vietnam (IPVN), industrial production of China (IPCN) and real exchange rate (RER). My data have over 140 observations and they are percentage change year over year. My question are:
1. How many observations should we have when running VECM?
2. When I run Johansen test, I have the results: Trace test indicates 3 cointegrating eqn(s) at the 0.05 level and Max-eigenvalue test indicates 1 cointegrating eqn(s) at the 0.05 level. Which test should I use to run VECM and what are the different?
3. Based on the Johansen test, can I conclude that there are the long run relationship between 3 dependent variables to TB?
4. Does Wald test use to test short run causality?
Thank you very much and I am looking forward to hearing from you soon.
Best regards!
Hi Tran, Can u please come to Hossain Academy Group in Facebook so that we can share there? If you let me know your email address I can invite.
Sayed Hossain Dear Sir, I already like your facebook and post my questions as well.
Thank you very much for you response.
Best regards!
Sayed Hossain
Dear Sir,
My email is: khanhcuong126@yahoo.com
Thank you very much!
Cuong Tran Quoc Khanh Please join hossain academy group in the facebook and make question. Follow the following link.@login.php?next=https%3A%2F%2Fwww.facebook.com%2Fgroups%2F841025519270747%2F843415589031740%2F%3Fcomment_id%3D843444675695498%26notif_t%3Dlike
Sayed Hossain
Dear Professor Sayed Hossain,
Would you please resend me your link because I cannot access this link.
Thank you very much!
Good day Prof. thank you so much for this teaching very interesting and explanatory. Am working on minimizing risk of precious metals. is it possible to use VAR model, VECM or which method would be most appropriate. Thanks
+Dare Jayeola VAR is not suitable for that to minimize risk. However, Thank you. I would like to invite you to join Hossain Academy Facebook for greater interaction about economics, finance and econometrics with me. Thank you Sayed Hossain from Hossain Academy. Please join below and post your question.facebook.com/groups/hossainacademy/
I have watched your video, it's so helpful for me. And can you make some video about panel data in eview. Thank you
I am on the way to prepare model about Panel Data
Dear Sir thank you very much
I have a question i want to know the relationship between credit and market and i have some control variables like EPS ratio, loan growth rate, total assets, return on assets. My question is do all variables have to be non stationary to run VECM or i can ignore control variables stationarity and focus on market and credit risk only??
Heba Mohamed In VECM model all variables should be non stationary at level but after first differenced they should be stationary. Are u using EVIEWS? or other software?
E views however, i am having trouble to set my model although i completely understand your video. May i tell you my variables and model may be you can help me understand how to run it...thank you sir
Sayed Hossain For VEM model you can see two videos to understand the whole concept. I am taking it from Hossain Academy below.
Sayed Hossain th-cam.com/video/MSLNkgyygP0/w-d-xo.html
I watched them twice. my problem is that not all of my variables are non stationary on level. second i got 12 banks from, china, Malaysia, South Africa, Egypt, Indonesia, and Brazil.
What i want to do is to see if credit risk and market risk are affected by each other and my control variables are:
GDP Growth rate
CPI
Return on Assets
Total Assets
Efficiency Ratio
Loan growth rate
Earnings per share
I am flexible in dropping any of my control variables.
Is running an unrestricted VAR as shown in previous videos suitable for me if they are not all non stationary at level?
Thank you very much you are a life saver cannot be thankful enough
AOA Sir; i have 9 variables in my model including dependent variable. If C(1)is negative and probability is significant at 5% level by checking long run causality but overall probability of overall model is not significant at 5% level then what should we do for making our model significant?
Thank you
Dear Mr Hossain, if I want to test the dependence of D(Y) on X(-1) (and not D(X(-1)) within VECM, how can I do it in EViews? Thank you in advance
Dear Sir Sayed.
your lectures are very helpful for me. i must say thanks to you. here i have some issue that in my model i take 3 lags and there is existence of long-run and also short run causality. what should i do ?
It can happy that b/c in short you are using an other test called wald test
Dear sir Thank you so much for such a nice delivery of content.
I have a question that it C(1) is -ve and insignificant than what would be our conclusion about long run causality
Thank you. I would like to invite you to join Hossain Academy Facebook at below link and post your question there. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Hi Neeru did you find the answer to this?
In that case there would be long-run causality from Exp, gdp to cons.
on the one hand the johnson test suggest there is long run relationship among variables, on the other hand, we find no casuality both in the short run and long run. How are these two results compatible??
Thank you. I would like to invite you to join Hossain Academy Facebook Group at below link and join our group discussion. Thank you. Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Sorry, i said when C(1) is significant, but is has positive sign. how do we conclude?
Anicet Irabaruta Hi....Please join our Hossain Academy group in the facebook below.facebook.com/groups/841025519270747/
I have 7 variable which 2 of them non stationer in level but stationer in 1st difference... can I use VAR? because there is condition for using VAR which is there is no cointegration
Houssain,i 'm glad to your lesson is running too well...but i need some detail about dummy or qualitatives variables
+jules karl You can put dummy in the exogenous variable box below.
In VECM, all variables are I(1). I got rank(2). but when comes to the VEC equation, I got many insignificant results if I keep all variables like that ( meaning that state turns into D1). when change into D.varname ( and stata turns into D2.varname) I got amazingly good results. what to do? keep varnames or change into D.varname before running VECM. thanks
Zoundi Zakaria Please join Hossain Academy Facebook below and post your question. Indeed I am there to share with you. Thank you Sayed Hossain from Hossain Academyfacebook.com/groups/hossainacademy/
Dear Sir.
All your videos are crucial and easy to understand. Thanks for it
My paper use Eview for running the long-run relationship betwen Exchange Rate anh Stock index with considering Macro index (CPI, unemployment). I had used all test ADF, Granger and Johenson. With all your Video i knew how to do and got the result the same what you did it. I want to thank you alot.
However, i have stucked with final step, using Nonparametric Cointergrating regression. Do you know it and how to run it. May you guide me how to run if you know.
Thanks Sir
Jane
Le Jane Hi Jane, please join Hossain Academy facebook group where we can discuss. Thanks. See below link.facebook.com/login.php?next=https%3A%2F%2Fwww.facebook.com%2Fgroups%2F841025519270747%2F843415589031740%2F%3Fcomment_id%3D843444675695498%26notif_t%3Dlike
Sir now when we do not have a log run and short run relationship then this means that there is no cointegration and we must do VAR analysis?
+sandeep pandey
Yes VAR is the answer.
Sir, your lessons are very good, can you please upload video on Non linear ARDL in stata. thanx
+Naveed Raza Not yet done. However, I would like to invite you to join Hossain Academy Facebook for greater interaction about economics, finance and econometrics with me. Thank you Sayed Hossain from Hossain Academy. Please join below and post your question.facebook.com/groups/hossainacademy/
sir, when i try to run vecm ...it shows insufficient number ofobservations.i hv data of fdi and 4 more variables annual data from 1991-2018. kindly guide
Dear. if the residuals are serially correlated and doesn't desirable as you have done in the video, wt shall we do?
+Henok Fasil
If there is serial correlation model is not acceprable at all.
what happen if the cointegration number is different for trace test (7) and max eigenvalue (4)? which of the value should i choose to run the program? 4 or 7?
Dear Gunadi, I would like to invite you to join Hossain Academy Facebook at below link and post this question there, I shall respond. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
if error correction term positive and significant
what if C(1) has a positive sign but significance? Thank you sir.
Means there is no long run cause
Hello! I'm trying to set up a VECM with CDS time series.
Unfortunately, all coefficients come up as "NA". Any idea what the problem is? Thank you
When there is error in setting your data, then EVIEWS gives this type of result so far I have seen..
Sayed Hossain Solved it, thank you (If someone else has the same problem: eviews wasn't opening my time series properly for some reason. A restart solved it)
Another quick question if you have the time: I have 3 time series, and the Johansen test (trace test) indicates 3 cointegrating equations.
However, when setting up the VECM, Eviews does not permit cointegration rank 3, only 1 or 2.
Must the cointegrating rank in the VECM necessarily be lower than the number of variables? Thank you.
sorry, if the model have trend (not to chosen 3, chosen 5(intercept and trend in CE-linear trend in VAR)), the order by veraible is not working. error: illegal date :) How can i understand my veriable is signifficant?
I guess your data setting was not correct or somewhere some problem. Go trial and error process.
How do I remove autocorrelation from VECM?
Mr. Hossain, please i want to know when i shall use Engle-Granger or Johansen? And why do you use OLS after running VECM, it isn`t enought to get the VECM?? So basically the process is: Test unit root, cointegration test,VAR, VECM and OLS? ... Please answer me it`s important, thank you!
Thank you. I would like to invite you to join Hossain Academy Facebook Group (Data Analysis) at below link and join our group discussion about modelling. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
my all the projected variables are stationary at first difference and run johanson co integration test where the trace test and eigan value test both are indicating 4 co integration equations. when i run the vecm where i slected the 4 co integration equation but the results are not producing it is giving error that co integration eq selection is not correct.
what is the problem here? kindly guide.
Thank you. I would like to invite you to join Hossain Academy Facebook Group at below link and join our group discussion. Thank you. Sayed Hossain from Hossain Academy.
What can we do for cointegration test if we don't have the same degree of integration??
Then you can not run johansen cointegration test, so can not run VECM also. So run only VAR model.
ok, thanks