(EViews10): VECM and 3-Ways Causality Checks (2)

[@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, support my Channel with your subscription and sharing my videos with your cohorts.

[@TheEnyoy]

Hey prof. dr. Adeleye, I want to know how to test for causality for the EC models?

[@pramodpandey8316]

Very simply explained a very effective interpretation. Thanks a lot

[@CrunchEconometrix]

Thanks for the positive feedback, Pramod. Deeply appreciated! 💕 May I know from where (location) you are reaching me?

[@asfiabinteosman5303]

Your explanation is very easy to understand. Please carry on The good job

[@CrunchEconometrix]

Glad to hear, Asifa...thanks!

[@ladymelodi]

I appreciate your efforts thank you very much.

[@CrunchEconometrix]

Glad you like them, Oyun...thanks!

[@NehaRajput-tv5op]

how to determine the direction(bidirectional, unidirectional) of causality through VECM

[@CrunchEconometrix]

Hi Neha, that is what this video and others explains.

[@NehaRajput-tv5op]

Do we need to determine the vecm model for each variable seperatly as i get different results in finding causality between revenue and expenditure if i take total revenue as dependent variable at one time and total expenditure as dependent variable at the other time

[@CrunchEconometrix]

No.

[@chandnirana369]

Thank you ma'am 😊

[@CrunchEconometrix]

You are welcome, Deepak 🥰🙏

[@patsonshikaaba6357]

Greeting from Zambia

[@CrunchEconometrix]

Same here from Nigeria...keep watching, keep sharing!

[@ti_jaheed]

Thanks from Bangladesh

[@CrunchEconometrix]

Awesome, Tj! Kindly share my videos with your students and colleagues in Bangladesh...thanks!

[@ntirabampadesire8712]

thank u very much. this was so helpful

[@CrunchEconometrix]

U're very welcome, Nti! I'll appreciate if you can please share my videos with your academic community...thanks!

[@kevinwilliamtimothy3545]

Hi Dr! thanks again for the explanation. I have question on the long run causal effect. if I understand it correctly from your explanation, the ECT/lambda is used to determine the long run causal effect of all variables towards 1 selected target variables. can we test long run causal effect by variables to the DV?

[@CrunchEconometrix]

That is information provided by the ECT/lambda.

[@doyinsolagbemisola8940]

thank you so much ma. please, what is the rule of thumb for rejecting the null hypothesis in the long-run causal effect, (for example, reject null when p value is less than or equal to ... or reject null when p value is greater than or equal to ...) thank you ma.

[@CrunchEconometrix]

Hi Doyin, once the pvalue is statistically significant at the 1%, 5%, and 10% the null hypothesis of no causal relation is rejected.

[@doyinsolagbemisola8940]

@@CrunchEconometrix thank you ma.

[@dylan2859]

@CrunchEconometrix Sir, from the regression above, what would be the estimated model equation with only significant terms?

[@CrunchEconometrix]

Hi Dylan, I am not clear about your query? Kindly recast.

[@ladymelodi]

In the long run you reverse the signs while interpreting. What about the signs in the short run below?

[@CrunchEconometrix]

Oyun, you are mixing things up. Signs are not reversed in VECM but in Johansen Normalisation Equation.

[@delphinehounye4328]

Hi! Thanks for your work! I have a question: How to get vecm long run p-value ? I know how to get for the short run. So, thanks for helping!

[@CrunchEconometrix]

Hi Delphine, use the t-stat.

[@delphinehounye4328]

@@CrunchEconometrix Thanks! Very useful!

[@samfisher1250]

what does it mean if i have only 2 variables which is farmgate and retail. i used retail as the left hand variable. the ect for retail is significant and negative but the ect for farmgate is significant but non negative. how do i interpret this?

[@CrunchEconometrix]

Hi Sam, your query is unclear. Kindly rephrase. Thanks

[@jananiravinag]

What is the implication of r squared in VECM?

[@CrunchEconometrix]

R2 follows the same interpretation in econometrics.

[@jananiravinag]

@@CrunchEconometrix If the R squared value is below .60, how to go forward? Ur channel is just right for newbees like me. Thanks a ton and more power to u!

[@CrunchEconometrix]

Jananiaravinag, you may need to add explanatory variables to improve the goodness-of-fit of the model.

[@poojakumar4819]

Thank you for all the videos, they have been very useful for my thesis. I have a question, how would you interpret the result if the independent variable is not in form of log, for example if it was PCE and not LNPCE? - How would you interpret the PCE in relation to LNPDI ?

[@CrunchEconometrix]

Hi Pooja, here is the link to my blog that addresses your query cruncheconometrix.blogspot.com/2018_01_21_archive.html. Go to Model 3. Thanks.

[@lincoln10]

Thank you very much. I have a question. @7:55 there is a long-run causal relationship between the variables but which variable cause pce in the long-run? gdp or pdi or jointly?

[@CrunchEconometrix]

Hi Lincoln, if you follow, you can easily deduce which variable causes pce and whether it is individually or jointly. I was quite clear in explaining the results. Thanks.

[@kamleshpahurkar5966]

@@CrunchEconometrix Can you please make it more clear, @7:55 there is a long-run causal relationship between the variables but which variable cause pce in the long-run? gdp or pdi or jointly? We say if ECM term is significant there is a long-run causal relationship but that is between which variables as if in equations there are 3 variables not only 2

[@omogbehinremilekunn]

Ma, please how do i interprete my VECM when i have 2 cointegrating equation? Do i just interprete the first cointegrating equation

[@CrunchEconometrix]

Hi Omogbehin, I explained in my Johansen Cointegration video that you keep-it-simple by using one CE which is the one normalized on the outcome variable.

[@ethancheng7565]

Hi Dr, thank you for your videos, they do help me a lot. I have a question about the short run casual effect analysis. Simply, in your video, C(3) only correspond to DLnPCE(-1) so that you can only analyze the P value of C(3). But what if C(3) correspond to DLnPCE(-1) with prob 0.15 but C(4) correspond to DLnPCE(-2) with prob 0.00? (Assume my optimal lag is 3)

[@CrunchEconometrix]

Hi Ethan, thanks for the encouraging feedback. Deeply appreciated! Watch my Causality videos again. Causality is between depvar and explvars. If pvalue is not significant implies zero causality. So, you need to read textbooks and articles to understand pvalues.

[@sibykm]

Madam, do we need to revert the sign of short-run coefficients while interpreting them?

[@CrunchEconometrix]

No Siby, only the results of the Johansen Normalization Equation are reversed during interpretation. You interpret the short-run results directly.

[@ibiagolikaibinabo5571]

Port Harcourt , Rivers State, Nigeria.

[@CrunchEconometrix]

Thanks Ibia. I'll appreciate if you can share the link to my KZbin Channel with your colleagues. They need to know... Thanks!

[@kokobeseyoum2040]

Thank you very much for your best analysis. My question is how we define the sign of short term causality? Do we still change the negative coefficient to positive and positive to negative when we interpret?

[@CrunchEconometrix]

Compliment is humbly taken, Kokobe. Simply follow my interpretation. Causality does not factor in the signs only the statistical significance. For instance, if coefficient of X on Y is -0.087 and significant at 5% level, it simply shows that there is a uni-directional causality from X to Y. May I know from where (location) you are reaching me?

[@siya7793]

thank you so much for making these videos. Really appreciated. I have one small question, what if the coefficient of my target ECM has a positive sign (30 observation, 7 variables), how to interpret it or is my model wrong?

[@CrunchEconometrix]

Positive ECM imply that the model is explosive and diverges in the long-run. No convergence to equilibrium. You may need to change the regressors and re-estimate.

[@at4859]

Hi greeting from Indonesia i am so happy finding your KZbin channel while i am looking for some references about causality on vecm. It helps alot I like your style which is fast but still easy to understand. if you don't mind to answer, i have a question related to this video. you told me that we can find long run causality by looking at ect, if it is negative and significant so it has causality. but can we split the causality into two variables? as on your model there are 3 variables GDP PDI PCE. for the long run causality how can we say PDU cause GDP or PCD cause GDP?

[@CrunchEconometrix]

Hi Adi, thanks for the positive feedback, deeply appreciated! There are so many econometrics textbooks and published articles on causality test, you can reference any of them. Also, run a 2-variable model instead of 3 and find the causal relation like I did in the video. May I know from where (location) you are reaching me?

[@UsmanAli-kf5ji]

CrunchEconometrix Thanks for such a nice video explaining VECM procedure in a well manner. I need to ask only one question about "Strong Causality". I saw many papers on this and I have also utilized this strong causality concept in my papers. But I do not know the concept of how "ECT and regressors together" can define STRONG causality. can you please provide a little explanation and also some references please.

[@CrunchEconometrix]

Hi Usman, thanks for the kind comments...humbly taken. Regarding your query, my explanations on the types of causality subsists and it will simply amount to repetition going all over them again. But you can see Asongu et al (2015) - "Testing the Relationships between Energy Consumption, CO2 emissions and Economic Growth in 24 African Countries: a Panel ARDL Approach" for indepth explanation of "strong causality". Hope this is helpful and I'll appreciate if you can share my videos with your students and academic community....gracias!

[@UsmanAli-kf5ji]

@@CrunchEconometrix Thanks for your response. I'll consult Asongu et al. (2015). Thanks again for sharing the reference.

[@nomaswazi_tshabalala]

Good Day Dr, can I use this method for panel data?

[@CrunchEconometrix]

Not sure. I have only applied it to time series VAR.

[@ghadaommezzine3065]

How can i interpret the significativity of my variable by the result of test VECM ? other , i can't interpret the sgnificativity of independents variables according to this method ? what's the interpretation of non significative impact of the ECT according to the result of test VECM ?

[@CrunchEconometrix]

Hi Ghada, pardon my very late response. Please watch my VECM videos and adapt the interpretation given. Thanks.

[@mkjoshi21]

Madam, in the VECM dialog box in Eviews, why we need to keep 1 in the cointegrating equation , even though we may get more than one cointegrating equation using Johansen Cointegration Test.

[@CrunchEconometrix]

I explained why in the video on JCT. Kindly watch it. Thanks.

[@ahmadz113]

Hello Dear , you are great as always . One question , If i want to run granger causality test between two variables . One is stationary at level and one is stationary at first level . Shall i take first difference for both of them or only one ?

[@CrunchEconometrix]

Hi Ahmad, you cannot run this type of Granger test in VAR/VECM be variables have mixed order of integration. Use ARDL. watch my CAUSALITY videos on how to proceed.

[@ahmadz113]

@@CrunchEconometrix thanks but what if I have variables which are stationary a mix of 2nd difference, 1st difference and level . What kind of causality check i can run ? And what kind of forecasting model ?

[@CrunchEconometrix]

Toda-Yamamoto procedure for the causality. You may need to check other online resources for more information on how forecasting.

[@ahmadz113]

@@CrunchEconometrix any recommendation where to check for forecasting ?

[@ibiagolikaibinabo5571]

Good work madam, I have accessed some of your videos and say you have doing a great job. Please I am interested in accessing all you have in E-views possibly from simple regression, descriptive statistics, unit root, co integration, vecm, var, bounds test, and other necessary test to diagnose test if I can access all of your packages I will be greatful.

[@CrunchEconometrix]

Thanks for the encouraging feedback, Ibiagolika. Deeply appreciated! Kindly browse through the EViews Playlist for videos related to your interests. Please may I know from where (location) you are reaching me?

[@anupamsabharwal4385]

when i am running VECM eviews showing error of insufficient number of observations. kindly guide me

[@CrunchEconometrix]

Hi Anupam, it is an indication that you do not have enough observations for at least one of the variables to accommodate the lag structure.

[@Davidsiregar88899]

So if the coefficient is negative we interpret this as positive?

[@Davidsiregar88899]

is the interpretation of long run and short run different? because i already saw at your previous video and you interpret the short run different than the long run

[@Davidsiregar88899]

please help me madam

[@CrunchEconometrix]

Christian, I never said so. Please watch the video again.

[@CrunchEconometrix]

I gave clear interpretation about the long- and short-run. You may want to watch the video again.

[@CrunchEconometrix]

I have responded to your queries.

[@limwenhao1999]

Hi madam. I tried VECM and i have 3 cointegrating equation. Would you recommend using 1 or 3 for VECM. Because the result is different in the coefficients when i do OLS on the proc equarion.

[@CrunchEconometrix]

Hi Hao, kindly watch my video on Johansen Cointegration Test after which you watch those on VAR and VECM to understand what you need to do correctly. Regards.

[@limwenhao1999]

@@CrunchEconometrix yes I have watched. You suggest using just one cointegrating equation even if there is more than 1. But the results on the error correction coefficient are different.

[@CrunchEconometrix]

You should get different results when different approaches are used. Stick to one.

[@souayehsirine1376]

Hello miss i'm working on granger causality test (pairwise granger causality test) between exchange rate and international reserves so I found a unidirectional relationship from the exchange rate to the international reserves only at lag 8 how can we interpret it? thank you in advance

[@CrunchEconometrix]

Hi Sirine, Granger causality is performed on the joint lags and NOT individual lags of the variables. Watch my videos on the procedure again and follow/adapt the interpretations. Thanks.

[@anupamsabharwal4385]

hello mam, whenever i calculate VECM , it shows insufficient number of observations. kindly guide me. johansen countegration showing 5 cointegrated equations.

[@CrunchEconometrix]

Hi Anupam, that implies you have TOO many variables. I also mentioned in my video on JCT that it is ADVISABLE to use only ONE cointegrating equation even if the result indicate more than one. Thanks.

[@Yasin-bq5pn]

Dr. why i should decrease the lag by one? in my analysis, the var says to use 2 lags. however, when i use 2 lags, and the variables are 1(1) and co-integrated, there is no causality among the variables at all; but when i reduce the lag to 1 lag, there is bidirectional causality running from one independent variable to the DV. could you please enlighten me?

[@CrunchEconometrix]

Hi Yasin, the reason why you have to drop a lag has been explained in my VECM video. VECM is simply a VAR in 1st diff which means you lose a lag during specification. I do not understand the rest of your queries. Are you estimating a VAR model or VECM?

[@Yasin-bq5pn]

@@CrunchEconometrix Thank you Dr. for the feedback. i runned ARDL model, and the variables are 1(1) and co-integrated, but i'm confused of what kind of test i should run for the causality test. is it Granger Causality through VECM?

[@CrunchEconometrix]

Watch my videos on "ARDL and Causality Checks"

[@ibiagolikaibinabo5571]

Good evening, lam Ibinabo from Port Harcourt. I want to know if I can get access to an Online diploma programme with you on econometrics.

[@CrunchEconometrix]

Hi Ibinabo, I am not accredited to run a Diploma program on my KZbin Channel. Hopefully, I will. Someday. Thanks 😊

[@ifedolaolabisi3015]

Dr, do the joint causality have same meaning as long run causality ?

[@CrunchEconometrix]

Hi Ife, I'll say so since it refers to the regressors (Xs) influencing movement in the outcome variable (Y) in the long-run. Referring to the ECT.

[@ifedolaolabisi3015]

@@CrunchEconometrix Alright, thank you.

[@CrunchEconometrix]

@@ifedolaolabisi3015 U're welcome...kindly share my KZbin Channel link with your students and academic community. Thanks! 💕

[@meditatewithme4933]

hi Dr, it's me again. may i know why is the long run coefficient intepreted inversely here?

[@CrunchEconometrix]

Che, I gave the reason in my previous response to you and also watch my video on Johansen cointegration for detailed information. Thanks 😊

[@Wendy-kr7dr]

Hello ma'am, the significant value usually is lower than 5%. My result shows the p-value is 0.7. Can I set the critical value at 10%? Thanks

[@CrunchEconometrix]

Hi Wendy, you are mixing 0.7 (70%) with 0.07 (7%). If the second is the pvalue, then you can decide to set your alpha to 10%. Hope this helps.

[@Wendy-kr7dr]

@@CrunchEconometrix Thank you so much!! :)

[@Wendy-kr7dr]

@@CrunchEconometrix Hi ma'am, sorry for asking you again. As I have the long equation, so I Copy the first equation, and apart from the lambda of dependent variable on C(1), can I use the lambda before D(other variables) to interpret other variables' casualty relationships?

[@CrunchEconometrix]

@@Wendy-kr7dr I have no idea what you are talking about but follow my guide and adapt the interpretation.