(Stata13): VECM and 3-Ways Causality Checks (2)

A statement such as “X causes Y” will have the following meaning in different scenarios and disciplines such as X leads Y, X is the only cause of Y, X is only one of the possible causes of Y, X must always lead to Y (that is X determines Y), the occurrence of X makes the occurrence of Y more probable, X is a probabilistic cause of Y, X must occur either before or simultaneously with Y, but not afterwards, past values of X forecasts future values of Y. But Regression analysis deals with the dependence of one variable on other variables, it does not necessarily imply causation. In other words, the existence of a relationship between variables does not prove causality or the direction of influence. But in regressions involving time series data, the situation may be somewhat different. Short-run causal effects: through the F-statistics and the statistical significance of the regressors. Long-run causal effects: through the statistical significance error-correction term (applicable to VECM only). Joint causal effects: through the F-statistics and the significance of the independent variables and the statistical significance error-correction term (applicable to VECM only). Unidirectional causality: occurs from X to Y if the set of estimated coefficients of the lagged X are significantly different from zero and the set of estimated coefficients of lagged Y are not significantly different from zero. Bi-directional causality: occurs from X to Y if the set of estimated coefficients of the lagged X are significantly different from zero and vice-versa. Independence: occurs from Y (X) to X (Y) if the set of estimated coefficients of the lagged Y (X) are not significantly different from zero. Using Stata13, this video shows you how to perform causality tests in three different ways within a VECM framework and interpret the results.
Here is the link to the ex21-1.wf1 dataset (EViews file) used for this tutorial (endeavour to have a Google account for easy accessibility): drive.google.c...
Follow up with soft-notes and updates from CrunchEconometrix:
Website: cruncheconometr...
Blog: cruncheconomet...
Forum: cruncheconometr...
Facebook: / cruncheconometrix
KZbin Custom URL: / cruncheconometrix
Stata Videos Playlist: • (Stata13):Estimate and...
EViews Videos Playlist: • (EViews10):Interpret V...

[@g.hiwotmolla2929]

extremely so good and help full and easy to understand in VAR and VECM thanks too

[@CrunchEconometrix]

Thanks for the encouraging feedback, Hiwot!

[@toluchristiana9565]

Hello Ma, I just wanted to drop an appreciation message. Your videos helped me a lot during my masters dissertation. Your efforts and time is greatly appreciated and recognised, please continue this great work! God blessings’!

[@CrunchEconometrix]

Good to know Tolu, your feedback is very encouraging and appreciated. It's the real essence of setting up the YT Channel. I wish you well in your future endeavours....and kindly tell others about my Channel.

[@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.

[@surojitdey574]

All your videos are very informative.

[@CrunchEconometrix]

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

[@yanuozhou6028]

Dear sir, when concluding short term causality effects in your video, you are referring to the F statistics of the Wald Test, right?

[@CrunchEconometrix]

Yes, Yanuo...and I'm sure I explained that.

[@erictiyabe5985]

Thank you very much for your interpretation, my question is, What if the _ce1 (adjustment parameter) is not significant at the D_lnpdi, which is the dependent variable, does that mean there`s a problem with the model?

[@CrunchEconometrix]

Not exactly. It only shows that there's no adjustment to long run equilibrium. If you change some of your independent variables and/or lag structure, you may get a different outcome.

[@erictiyabe5985]

@@CrunchEconometrix okay thank you 🙏🏽

[@dgscholar]

Hi there, in this video, is it correct to say PCE has long run causal effect on PDI (since PDI is the dependent variable) because the ECT of PCE is significant?

[@dgscholar]

Another question is how would you interpret a positive ECT?

[@CrunchEconometrix]

Yes

[@CrunchEconometrix]

@@dgscholar Positive ECT means no convergence to long-run equilibrium. The model is explosive.

[@EdgarFaustino13]

what if the lag criteria is 1 lag? can I use a VEC model? in minute 3:13 you say that maximum ag should be k-1. Thank you!!

[@CrunchEconometrix]

Edgar, in that case retain the one lag since VECM cannot be a static model.

[@EdgarFaustino13]

@@CrunchEconometrix Thank you!

[@kelv2629]

Hello CrunchEconometrix, if the error correction term is positive and significant at 10%. Does that mean that there is a long-run causal effect but do not converge in the long run equilibrium? Or if it is positive and significant, there will be no causal effect in the long run?

[@CrunchEconometrix]

Kelv, positive ECT implies DIVERGENCE in the long-run.

[@elshanalakbarli4966]

Dear Madam, Thanks a lot for your videos! It has been very helpful for me as a starter in STATA. However I have a question regarding the results of Johansen normalization restriction imposed table. Why do you interpret the results in reverse? Is there a specific reason for that? Thanks in advance!

[@CrunchEconometrix]

Hi Elshan, because the Johansen Normalization equation is in IMPLICIT form. So, you need to interpret the results in reverse (EXPLICIT).

[@fatihaelagri7753]

hello, if we have done an ardl modelling and we check the cointegration how to do the causality test; do we have to estimate a var or vecm equation?

[@CrunchEconometrix]

Hi Fatiha, kindly watch my ARCH-Causality videos on the steps required. Thanks.

[@kamleshpahurkar5966]

if I want to check for the long-run causality between LPDI and LPCE how will check?

[@CrunchEconometrix]

Kamiesh, kindly watch the foundation video on causality. Answers your question.

[@cronanryan5066]

Hi CrunchEconometris! Love your content! Im just curious whether or not you are able to create IRF/OIRFS'S by estimating the results from a VEC Model rather than varbasic. Thanks.

[@CrunchEconometrix]

Thanks Cronan, for the encouraging feedback. Not sure if you can do that from VEC.

[@fredli2888]

Hello professor, thanks for the amazing content. Just a quick question, is the coefficient of ce shown in the short-run result table lamda? Thanks in advance!

[@CrunchEconometrix]

Hi Fred, yes. That is the speed of adjustment.

[@himanisharma5417]

Hi, thank you for the great work! Your videos are really informative. I was facing a question regarding the difference between vecm and the ecm model. The vecm model is used if all equations show cointegration. However, if only one of the three questions shows cointegartion and the the other two do not, we will estimate the one that shows cointegartion with the ecm instead of vecm? How do we do that?

[@CrunchEconometrix]

Hi Himani, thanks for the encouraging feedback. Deeply appreciated! You are mixing up the information. ECM relates to single-equation models while VECM relates to a system of equations. I only constructed an ARDL-VECM for illustration purposes only. Watch my videos on VAR/VECM for better understanding. Thanks

[@khouloudbradai3731]

how can I estimate the causality in the long term, I mean after estimate VECM what kind of test shall I do to interpret whether there's causality between the variables or not? thanks, in advance ^^

[@CrunchEconometrix]

Hi Khouloud, this video and other related ones explains Causality in VECM. What else do you need?

[@khouloudbradai3731]

@@CrunchEconometrixI understand that for the short term we estimate the VAR model then we use the causality granger test. my question is after estimating VECM models is there any test after to interpret the causality?

[@CrunchEconometrix]

Kindly watch my VECM Causality videos. Well explained. Thanks.

[@khouloudbradai3731]

@@CrunchEconometrix ok , thanks a lot

[@haotianfeng1163]

Hello, I want to ask one question that if the lag length is not 1 like the example you offer, is 2 or more, how should we judge the casual relationship between variables?

[@haotianfeng1163]

Especially how to judge short-run casual relationship

[@CrunchEconometrix]

@@haotianfeng1163 Causality test is performed on the variable regardless of the number of lags.

[@CrunchEconometrix]

@@haotianfeng1163 I explained short-run causality. Kindly watch the series again. Thanks 😊.