Two brief questions: 1) What does it mean to assume the variables are 'stationary' and how do we test for that? 2) Can you provide a basic interpretation of your results in terms of 2,4 and 5 lags? The results were insignificant at 2 and 4 lags, but significant for 5 lags (LR > CONS). Do we interpret this as: LR does Granger cause CONS, but it is a delayed effect (5 years)?
Sayed, many thanks for this and the other videos on Granger. I'm very please you had a worked example and went through it in eVIEWS. A lot of others have glossed over the acception criteria but you spent time reinforcing it for each of the results, which was just what I was looking for. 51stState
Normally if the F statistics is found significant (P value less than 5 percent) it means that all the independent variables jointly can influence dependent variable.
Dear Sayed, I really appreciate your efforts and knowledge in econometric. I am doing Granger Causality analysis between two variables GDP and Investment of Sweden and I want to ask three little questions about my empirical results. Q#1 Both the variables are non stationary and their 1st difference is still have unit root so I used 2nd difference of both the variables. So is it ok to use series of diffidence twice or integrated of order 2 in Granger test???
Videos on E views are really beneficial for those who have thrust for knowledge.Excellent Work.Will the videos on Panel data be available on your Site. God Bless You
Stationary Test actually unit root testing. There is a video in my channel. You can check it. Yes, changes of lag affect the result all the time. But you have to find out the optimal lag to be used in the model.
Thank you so much Professor Sayed Hossain. I appreciate your teachings on e-view a lot. you have blessed many with it. Please how can I interpret causal relationship among variables from an ECM result?
Dear Ragiland, 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/
Thank you very much Sayed; I will find and watch your stationary test video. Am I interpreting the lag results properly? LR effects CONS, but in a delayed manner.
I have already uploaded Granger Test video in my website. All the explanations are there. Please see it. Go to Sayed Hossain Website in google and can see my full website
The acceptance/rejection of hypothesis may vary on the significant level we are choosing.In that case what is the typical significant value we should use in a practical example? Also if we are explaining the rejection/acceptance what is the best way to explain in non statistical means? (is it as we are 95% confident that LR causes CONS likewise?)
Dear S.Hossain. As on your great videos I have done the VECM tests which show long and short-run causality. Why do I also need to do the Granger Causality test? What is the difference? Many happy kind regards, Jane x
Thank you.I would like to invite you to join Hossain Academy Facebook at below link to discuss about data analysis, econometrics and statistical models. facebook.com/groups/hossainacademy/
I did not make videos with ARDL yet. But guideline is, if the variables are cointegrated, then you should run VECM and then check the causality using Wald test
Sayed Sir, I have a couple of questions please. I have two datasets spanning 132 months from January 2002 until December 2012, in intervals of one month and another spanning 108 months from January 2004 until December 2012. I carried out the Akaike information criterion (AIC) tests to find the optimum lags however I also computed all possible Lags (until eViews returned null’s and couldn’t compute granger anymore) When carrying out granger I get quite a few “Bi-directional causality” (i.e. 2 x Cannot Reject Null / 2 x Reject Null, Accept Alternative Hypothesis) in the 108 months set. In the 108 month dataset my last non Bi-Directional granger results are at 29th lag (so that’s almost 2.5years) you mentioned in a video that you wouldn’t use a lag over 10 is that based on a statistical rule, or a judgement call? Should I ignore the results at 29th lag? Two null hypothesis at 29th lag were; X does not Granger Cause Y , 79 obs , f=2.0342 , prob=0.0513 = 5.1% Y does not Granger Cause X , f = 2.0836 , prob = 0.0458 = 4.6%
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/
Sir, I have estimated an ARDL approach to cointegration. my variables are cointegrated. I want to study the causality between my variables now. Can I estimate Granger causality in VAR environment for my variables? or should I estimate VECM?
Dear. which video is talking about "weak exogeniety test of the engle"....i am looking to see my in VECM if there is weak exo. or not? please share me the link? thank you
my two variables are I (1). I want to study the Granger causality. it 's correct if I havn't use the test of cointegration and the estimation of VAR (or VECM).
One thing though your website link above is trying to go to HTTPS and fails! Best change it back to HTTP just incase it stops people finding your example files.
Respected Sir, I am doing Ph D. I have to apply the Granger Causality Test. I have two variables FII & SENSEX. I have collected their prices. Now what I supposed to do???How can I know whether data are stationary or not????will u pls help me and provide me the details what I should do with that data??? Seeking your prompt & positive reply sir. Thanking you,
granger causality is to show the direction causality between variables in which run?? long run or short run and there are many articles they use granger VECM models and they use granger for short run and VECM for long run why???
I'm sorry, I'm a beginner, what program did you use? could I use Stata 12 too? is there a way to determine the best number of lag periods? thanks in advance for your kindness
Q#2 If both the variables are not co integrated so we can not use VECM or both dont have long run association ship. Is it right to apply Granger test directly now to analyse causation??? Q# 3 IF both variables have causality at 5 lags (optimal lag according AIC) but don not have causation till 4th lag. So what kind of causality is at higher lags??? causality at more lags what actually tells about???is it strong or slight or what???comment on that plzzz
Dear Sir, Can you please elaborate on the Granger non-causality tests developed by Toda and Yamamoto (1995)? I actually dont know how to deal with it. I guess that will be a useful information.
dear sir, i'm doing a granger causality test of the 2 variables ; FDI and Total factor productivity base in one country consisting of 11 sectors for a period of 2000-2012.It is a panel data. I already input the data in Eviews base on balance panel data on 2000 2012*11.So ,Sir should the procedures be the same in running the granger causality as you have illustrated the above videos. thks!
Good after noon sayed ! I am a beginner and searching for the data for my research Does stock prices Granger cause GDP for this i need data of quaterly percent change in Nifty for the period 1995 to 2012. how can i get it. Kindly suggest me as i have tried it but could not succeed. thank you.
Dear Sany, Thank you. I would like to invite you to join Hossain Academy Facebook at below link and post your question there. If I know the answer I shall certainly respond. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
Hmmm… You tried several lags until you found “significance”. Once you did so, your chance of a Type I error exceeded 5%... which was arbitrary to begin with come to think of it.
+sarwan sandhu 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 u for uploading this video, it has been of great help...I have 5 Independent variables and 1 Dependent variables when I checked for granger causality between 2 variables I find that H0 is rejected. I mean x causes y and y causes x ...can u please let me know how to report this kind of result...
hi, I am using the causality test for testing for PPP. How do I comment the output? in which case I can say that PPP does hold? I am using exchange rate and Price level as variables. Thanks.
Thank you for your teaching, would you please answer my question?. I have 6 variables that 5 of them are nonstationary, and I have to find causal relationship between them, two by two, how can I use Granger causality?
hello professor, I want to find causal relationship between firm specific and industry variable and I have panel data. how to run granger causality test in panel setting.
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/
thank you a lot for making this video and your help .. I'm doing a project, i have 2 variables with yearly time series.. what lag do you think i should use? to be honest i have no idea what it means or does... :\
+Sayed Hossain 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/
Null Hypothesis: Obs F-Statistic Probability Decision INT does not Granger Cause INV 33 1.46943 0.24726 Accept INV does not Granger Cause INT 0.07791 0.92525 Accept INFL does not Granger Cause INV 33 0.03934 0.96148 Accept INV does not Granger Cause INFL 0.69968 0.50522 Accept GDP does not Granger Cause INV 33 NA NA Accept INV does not Granger Cause GDP NA NA Accept EXC does not Granger Cause INV 33 1.44150 0.25359 Accept INV does not Granger Cause EXC 0.00262 0.99739 Accept INFL does not Granger Cause INT 33 0.14717 0.86381 Accept INT does not Granger Cause INFL 0.21083 0.81119 Accept GDP does not Granger Cause INT 33 0.07805 0.92512 Accept INT does not Granger Cause GDP 1.46953 0.24724 Accept EXC does not Granger Cause INT 33 2.12962 0.13774 Accept INT does not Granger Cause EXC 3.39305 0.04792 Accept GDP does not Granger Cause INFL 33 0.69974 0.50520 Accept INFL does not Granger Cause GDP 0.03934 0.96147 Accept EXC does not Granger Cause INFL 33 1.15898 0.32841 Accept INFL does not Granger Cause EXC 1.00582 0.37858 Accept EXC does not Granger Cause GDP 33 1.44145 0.25361 Accept GDP does not Granger Cause EXC 0.00261 0.99740 Accept
![Summer Walker - Heart Of A Woman [Official Lyric Video]](http://i.ytimg.com/vi/SFddctbZzw8/mqdefault.jpg)
![Gucci Mane & Sexyy Red - You Don't Love Me [Official Music Video]](http://i.ytimg.com/vi/j7IBqyFKT4c/mqdefault.jpg)
Two brief questions: 1) What does it mean to assume the variables are 'stationary' and how do we test for that? 2) Can you provide a basic interpretation of your results in terms of 2,4 and 5 lags? The results were insignificant at 2 and 4 lags, but significant for 5 lags (LR > CONS). Do we interpret this as: LR does Granger cause CONS, but it is a delayed effect (5 years)?
Sayed, many thanks for this and the other videos on Granger. I'm very please you had a worked example and went through it in eVIEWS. A lot of others have glossed over the acception criteria but you spent time reinforcing it for each of the results, which was just what I was looking for. 51stState
You are welcome
Normally if the F statistics is found significant (P value less than 5 percent) it means that all the independent variables jointly can influence dependent variable.
If you want to check whether the data is stationary or not, you need to do unit root testing. You can find two unit root testing in my videos
Dear Sayed, I really appreciate your efforts and knowledge in econometric.
I am doing Granger Causality analysis between two variables GDP and Investment of Sweden and I want to ask three little questions about my empirical results.
Q#1 Both the variables are non stationary and their 1st difference is still have unit root so I used 2nd difference of both the variables. So is it ok to use series of diffidence twice or integrated of order 2 in Granger test???
Videos on E views are really beneficial for those who have thrust for knowledge.Excellent Work.Will the videos on Panel data be available on your Site.
God Bless You
Stationary Test actually unit root testing. There is a video in my channel. You can check it. Yes, changes of lag affect the result all the time. But you have to find out the optimal lag to be used in the model.
May I ask how the optimum lag is determined?
You need to convert the variables into stationary first (even you require 2nd difference) and then you can run granger test
Thank you so much Professor Sayed Hossain. I appreciate your teachings on e-view a lot. you have blessed many with it. Please how can I interpret causal relationship among variables from an ECM result?
Dear Ragiland, 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/
My variable is stationary at first difference. Does it mean that i should create new series like Dx and then run Granger test?
Thank you very much Sayed; I will find and watch your stationary test video. Am I interpreting the lag results properly? LR effects CONS, but in a delayed manner.
if our variables are not stationary, is it ok to do granger causality test on their first differences?
That can always happen and that is why there is system equation such as VAR
thanks, this video helped me finish homework
I have already uploaded Granger Test video in my website. All the explanations are there. Please see it. Go to Sayed Hossain Website in google and can see my full website
The acceptance/rejection of hypothesis may vary on the significant level we are choosing.In that case what is the typical significant value we should use in a practical example? Also if we are explaining the rejection/acceptance what is the best way to explain in non statistical means? (is it as we are 95% confident that LR causes CONS likewise?)
Dear S.Hossain. As on your great videos I have done the VECM tests which show long and short-run causality. Why do I also need to do the Granger Causality test? What is the difference? Many happy kind regards, Jane x
thank you very much
Thanks! this was very useful.
Thank you.I would like to invite you to join Hossain Academy Facebook at below link to discuss about data analysis, econometrics and statistical models.
facebook.com/groups/hossainacademy/
I did not make videos with ARDL yet. But guideline is, if the variables are cointegrated, then you should run VECM and then check the causality using Wald test
Sayed Sir, I have a couple of questions please.
I have two datasets spanning 132 months from January 2002 until December 2012, in intervals of one month and another spanning 108 months from January 2004 until December 2012.
I carried out the Akaike information criterion (AIC) tests to find the optimum lags however I also computed all possible Lags (until eViews returned null’s and couldn’t compute granger anymore)
When carrying out granger I get quite a few “Bi-directional causality” (i.e. 2 x Cannot Reject Null / 2 x Reject Null, Accept Alternative Hypothesis) in the 108 months set.
In the 108 month dataset my last non Bi-Directional granger results are at 29th lag (so that’s almost 2.5years) you mentioned in a video that you wouldn’t use a lag over 10 is that based on a statistical rule, or a judgement call?
Should I ignore the results at 29th lag?
Two null hypothesis at 29th lag were;
X does not Granger Cause Y , 79 obs , f=2.0342 , prob=0.0513 = 5.1%
Y does not Granger Cause X , f = 2.0836 , prob = 0.0458 = 4.6%
Certainly you should not use 29 lags. Decide on the basis of 10 lag maiximum
very helpful, thank you so much.
Fantastic
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/
Dear Sir,
Regarding the lag selection, why did you should 2? Did you run a VAR to check the appropriate lag?
You are welcome
do you suppress the constant for these regressions? should you?
if after adding more lags there is no casuality then? please help i am stuck on it
Sir, I have estimated an ARDL approach to cointegration. my variables are cointegrated. I want to study the causality between my variables now. Can I estimate Granger causality in VAR environment for my variables? or should I estimate VECM?
Dear. which video is talking about "weak exogeniety test of the engle"....i am looking to see my in VECM if there is weak exo. or not? please share me the link? thank you
my two variables are I (1). I want to study the Granger causality. it 's correct if I havn't use the test of cointegration and the estimation of VAR (or VECM).
One thing though your website link above is trying to go to HTTPS and fails! Best change it back to HTTP just incase it stops people finding your example files.
I can not understand what actually the problem. But thanks for raising the problem.
Respected Sir,
I am doing Ph D. I have to apply the Granger Causality Test. I have two variables FII & SENSEX. I have collected their prices. Now what I supposed to do???How can I know whether data are stationary or not????will u pls help me and provide me the details what I should do with that data???
Seeking your prompt & positive reply sir.
Thanking you,
granger causality is to show the direction causality between variables in which run??
long run or short run
and there are many articles they use granger VECM models and they use granger for short run and VECM for long run why???
How can I use Granger causality in a VAR model? Is that the Granger causality/block exogeneity Wald test?
Happy to know it
Sir,would you tell me the stata command for granger causality tests in ARDL
?
Extremely slow video. Watching with 1.50 playback speed and yet it's extremely slow. Please make video length shorter and speak faster....
I'm sorry, I'm a beginner, what program did you use? could I use Stata 12 too?
is there a way to determine the best number of lag periods?
thanks in advance for your kindness
what are those?
Sir kindly comment on Questions 2 and 3 as well. I will be much obliged
It depends on your objective of study
thank you
Q#2 If both the variables are not co integrated so we can not use VECM or both dont have long run association ship. Is it right to apply Granger test directly now to analyse causation???
Q# 3 IF both variables have causality at 5 lags (optimal lag according AIC) but don not have causation till 4th lag. So what kind of causality is at higher lags??? causality at more lags what actually tells about???is it strong or slight or what???comment on that plzzz
I would like to know what if my dependent variable is causing the independent variable?
Dear Sir,
Can you please elaborate on the Granger non-causality tests developed by Toda and Yamamoto (1995)? I actually dont know how to deal with it. I guess that will be a useful information.
dear sir,
i'm doing a granger causality test of the 2 variables ; FDI and Total factor productivity base in one country consisting of 11 sectors for a period of 2000-2012.It is a panel data. I already input the data in Eviews base on balance panel data on 2000 2012*11.So ,Sir should the procedures be the same in running the granger causality as you have illustrated the above videos. thks!
Good after noon sayed ! I am a beginner and searching for the data for my research Does stock prices Granger cause GDP for this i need data of quaterly percent change in Nifty for the period 1995 to 2012. how can i get it. Kindly suggest me as i have tried it but could not succeed. thank you.
If I were in your position, I would have used DX (after first difference) data to run Granger causality.
Welcome. Panel data in future
Do you have any video is about Granger Causality test based on VECM model ?
Yes...it is there.
HEY GUYZ!, SO WHATS THE CRITERIA FOR LAG SELECTION?, IS THERE A SET OF RULES FOR LAG SELECTION?, OR DO I JUST SELECT WHICHEVER MAKES ME HAPPIEST?
sir, let me ask you one thing. Can Engle-granger only test for two variables?
Dear Sany, Thank you. I would like to invite you to join Hossain Academy Facebook at below link and post your question there. If I know the answer I shall certainly respond. Thank you once again, Sayed Hossain from Hossain Academy. facebook.com/groups/hossainacademy/
You can do it directly by selecting granger causality icon in Eviews.
Hmmm… You tried several lags until you found “significance”. Once you did so, your chance of a Type I error exceeded 5%... which was arbitrary to begin with come to think of it.
ihbarddx Yes you can add or deduct lags to get optimum result.
hi sir pls suggests me how to take zero co integration in vecm model, in vcm restrictions
+sarwan sandhu
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 u for uploading this video, it has been of great help...I have 5 Independent variables and 1 Dependent variables
when I checked for granger causality between 2 variables I find that H0 is rejected. I mean x causes y and y causes x ...can u please let me know how to report this kind of result...
What does F-statisitc tell us?
hi, I am using the causality test for testing for PPP. How do I comment the output? in which case I can say that PPP does hold? I am using exchange rate and Price level as variables.
Thanks.
Michele Izzo What is PPP?
Sayed Hossain Purchasing power parity theory
thank you sir
Personally I would choose stationary data to run Granger Causality..
Thank you for your teaching, would you please answer my question?. I have 6 variables that 5 of them are nonstationary, and I have to find causal relationship between them, two by two, how can I use Granger causality?
Najmeh Izadi So convert all variables into stationary and then run granger causality test...You need stationary data to run granger test,,,,
Sayed Hossain It means I should take first difference for all variables?
If needed then yes...make variables stationary first....
Sayed Hossain Can I use Toda Yamamoto causality test? Have you ever had any video about this test? Thank you so much in advance
Not yet I have
Normally it is the standard formula
I just took 2 lags without checking lag selection criterion...
it's short or long run direction??
Sorry your question is not clear to understand
hi please my variables where significant at one lag can i use the one lag period
+Angelina Ifurokienma It depends on your model objective
hello professor, I want to find causal relationship between firm specific and industry variable and I have panel data. how to run granger causality test in panel setting.
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/
i have my data on excel sheet...how do i insert them in e-views....im really lost....
It is simple job. You bring your curser on the EXCEL file and right click on the mouse and open your excel file under EVIEWS.
thank you a lot for making this video and your help .. I'm doing a project, i have 2 variables with yearly time series.. what lag do you think i should use? to be honest i have no idea what it means or does... :\
I have a video on Lag selection. Please see it.
will do. thanks!
sir are u from bangladesh?
Yes I am from Bangladesh
That's it!!!
please how can i explain when my result shows that savings does not granger cause interest and interest rate does not granger cause savings
+thomas indi
Post you results in Hossain Academy Facebook below.
+Sayed Hossain
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/
Null Hypothesis:
Obs
F-Statistic
Probability
Decision
INT does not Granger Cause INV
33
1.46943
0.24726
Accept
INV does not Granger Cause INT
0.07791
0.92525
Accept
INFL does not Granger Cause INV
33
0.03934
0.96148
Accept
INV does not Granger Cause INFL
0.69968
0.50522
Accept
GDP does not Granger Cause INV
33
NA
NA
Accept
INV does not Granger Cause GDP
NA
NA
Accept
EXC does not Granger Cause INV
33
1.44150
0.25359
Accept
INV does not Granger Cause EXC
0.00262
0.99739
Accept
INFL does not Granger Cause INT
33
0.14717
0.86381
Accept
INT does not Granger Cause INFL
0.21083
0.81119
Accept
GDP does not Granger Cause INT
33
0.07805
0.92512
Accept
INT does not Granger Cause GDP
1.46953
0.24724
Accept
EXC does not Granger Cause INT
33
2.12962
0.13774
Accept
INT does not Granger Cause EXC
3.39305
0.04792
Accept
GDP does not Granger Cause INFL
33
0.69974
0.50520
Accept
INFL does not Granger Cause GDP
0.03934
0.96147
Accept
EXC does not Granger Cause INFL
33
1.15898
0.32841
Accept
INFL does not Granger Cause EXC
1.00582
0.37858
Accept
EXC does not Granger Cause GDP
33
1.44145
0.25361
Accept
GDP does not Granger Cause EXC
0.00261
0.99740
Accept
+thomas indi Post it as picture
I have not done it also