VECM. Model Two. Part 2 of 3. STATA

[@shubhalalbanik7695]

Thanks a lot sir...I don't have any language to appreciate your help.

[@toromiguelangel]

Hello Sayed. Do you have any literature used for STATA to justify what you're saying about the long run causality found with the negative sign (and significance) of the ce1 coefficient inside the matrix? I believe you, but as I'm writing my undergrad dissertation I need to be able to cite that from some book or online source. Thanks.

[@amaranwar1249]

I understand this point but question is how to remove that?? A couple of more questions: VAR and VECM are the candidate estimation techniques for multiple regression time series. Is there any other technique? Also, what if the dependent variable is stationary at level 3 while independent variables are stationary at level 2 or 3. Can we still run VECM when the variables are cointegrated?

[@sayedhossain23]

If rank is 2 and if both error correction term is significant and contain negative sign I would say there is long run causality. Some say at least 1 out of 2 should be significant and negative sign to have long tun causality.

[@shubhalalbanik7695]

Sir, How can I know in VECM that either a coefficient at an individual lag of any independent variable has a positive effect or negative effect on the dependent variable . Moreover, how can I know the overall effect of all the lags of an independent variable on the dependent variable either it is negative or positive on the dependent variable.

[@sayedhossain23]

You are welcome

[@shubhalalbanik7695]

Sir, I mean that in VECM model we get two kinds of outcome one tells about the long run association of the variables and the others tell about short run effects. Now, If I use yearly data in my model, what does it mean by short run effect, will it be monthly effect or quarterly or yearly? I want to know the period of short run effect.

[@danieldelalingoh9379]

Hello Sayed, My thesis topic is modelling inflation, Interest rate, Exchange rate and money supply. An application of MGARCH model. No videos on mgarch? Any advise? I also would want to know if it makes sense that predictions made can be compared with VECM?

[@sayedhossain23]

there is already some videos related this issue. Please check.

[@sayedhossain23]

Just look at the coefficient of the independent variable. If it is positive, then affect positively, if negative then affect negatively the dependent variable.

[@sayedhossain23]

I have seen some literature that error correction term should be negative and significant. I have seen mostly in journal...You can see those on line.

[@Fredekkkkkk]

Hi, could you tell me why the error correction term must be negative in order for there to be long run causality?

[@tiffanyyuan9972]

Hi, Sayed, Thanks so much about this，would you Iike to give us a guildline when the rank is 2?

[@sayedhossain23]

So far I know there is no time period called short run effect. It is in the sort run means something like immediately.

[@shubhalalbanik7695]

Sir, I have found in a paper that if the Error Correction Term(-1) estimated coefficient is -0.87 (The estimated coefficient indicates that about 87 per cent of this disequilibrium is corrected between 1 year (if annual data)).So, If I use annual data in my VECM model, what will be the time span of short run effect of the coefficients of lagged variables ?

[@shubhalalbanik7695]

Sir, I have watched so many videos on VECM. I found that only the significance of the coefficient(s) is discussed. But the negative or positive effect of that coefficient(s) on the dependent variable is not discussed. I have devoted myself very firmly to find the video related to my problem but I failed to find one. Would you please help me find a video regarding my problem or would you please help me in any other way? I would be vary grateful to you for your help.

[@thomasrowe3959]

Hello Professor. When we have ran our VECM model and obtained the cointegrating equation. Do we have to reverse the coefficients in the normalised equation shown at the bottom of the screen? The coefficients do not match my expectations! Thanks

[@sayedhossain23]

+Thomas Rowe Yes you need to reverse the sign. Yes the coefficient may not match with expectation. Then you are in problem.

[@sayedhossain23]

+Sayed Hossain Dear Thomas, I would suggest you to join Hossain Academy Facebook at below link and post your question there. If I know the answer I shall respond. Thank you once again, Sayed Hossain from Hossain Academy facebook.com/groups/hossainacademy/

[@zhengzheng4873]

hi sir,what is mean "we are not happy with this model"? and what can we do next if "it is not a desirable model"?

[@sayedhossain23]

Zheng Zheng I am not happy as some of the statistical requirements are not met. In that case you have re-run the model again to fulfill all statistical requirements...

[@minyahilalemu4302]

concerning the vec post diagnostic test, you have no video on parameter stability. would you please help me, any how?

[@sayedhossain23]

+Minyahil Alemu Parameter stability you can use cusum test.

[@isaiahduterte]

I have 3 cointegrating equations but only ce1 is negative and significant. The others are only significant but positive. What do i do?

[@sayedhossain23]

+isaiah duterte When you run VECM model use 1CE not 3CE

[@sayedhossain23]

+Sayed Hossain I would suggest you to join Hossain Academy Facebook at below link and post your question there. If I know the answer I shall respond. Thank you once again, Sayed Hossain from Hossain Academy facebook.com/groups/hossainacademy/

[@sayedhossain23]

Hi I have not done anything with MGARCH model. So can not help now. Sorry

[@SnowRider757]

Hi, if you did VECM with rank 2 and lag 2 which in the table becomes LD1, but you have two error correction terms, which one do you use? Thank You

[@SnowRider757]

also how does one find the F statistic for the VECM?

[@sayedhossain23]

If there are 2 cointegrating equation (2 error terms) and if both of them have negative sign and signiiicant meaning that there is a long run causality running from independent to dependent variable.

[@shubhalalbanik7695]

Thank you sir

[@amaranwar1249]

When you say that model is not efficient because the presence of auto-correlation in the residuals or/and failure of normality in the residuals calculated through Jarque-Bera test, what is the solution when we already know that variables are cointegrated?

[@sayedhossain23]

You need to remove serial correlation from the model in all cases...

[@sayedhossain23]

I can not understand your question.

[@arfanbd]

Sir, I found your video lecture on VECM very useful. Thanks for your great effort. When i am running this VECM i got two problem. pls. give the answer of my questions. so that i can correctly run it. 1. My VECM is based on 4 variable. and two of them are found stationary after 2nd difference. in this case should i run VECM based of 2nd difference result or at level. 2. There is a two cointegration relation found in the JC test. In VECM i found one error-correction term is negative and another one is positive. in this case is the model correct and can i take only negative error-correction term to explain the model? pls. give your king replay.. Have a nice day

[@sayedhossain23]

Uzzaman Arfan 1. No you can run VECM. Before VECM you should run Johansen test 2. You can not run Johansen test here. To run Johansen test, all variables such be I(1). You check my Johansen Test videos

[@lijialin684]

That's why people go to university to avoid lullaby like this.

[@sayedhossain23]

Thank you. I would like to invite you to join Hossain Academy Facebook at below link and post your question there for feedback. Thank you, Sayed Hossain from Hossain Academy facebook.com/groups/hossainacademy/

[@sayedhossain23]

You are welcome