Stata Tutorial: Vector Auto-Regression in Stata

Stata commands used to specify and estimate a Vector Auto-regression model and generate Impulse Response Functions given a specific Cholesky Ordering.

[@toriyx]

Great summary Mike, thank you for creating this video!

[@npvlong]

Great video!! Thank you

[@deeptysarder6797]

Nice presentation

[@drjabirrahman]

Thank you!

[@noorrizvi1657]

Thank you.

[@rabihel-habta313]

amazing! could you also provide an explanation when it is better to go for VAR or just performing an OLS on time series data?

[@suzzzon]

Awesome! Thank you.

[@mikejonaseconometrics1886]

Glad you found it useful - let me know if there are other topics you'd like covered. Thanks!

[@chrismau8262]

@@mikejonaseconometrics1886 Hello sir could make a video about Local projection method Jorda(2005). Its an alternative to VAR in order to estimate cumulative irf

[@patriciocantu5587]

This was so useful, thank you so much. By the way, if anyone is getting error 603 constantly when trying to create the irf graphs, I found the following command to be useful, hope it helps. Not sure what it does but it helped! cd `c(tmpdir)'

[@lucafavero6291]

Dear mike, thanks a lot for this video. I didn't fully get what is the conceptual difference between the effect we capture with irf and the one we get with var

[@carssurfer]

Hello, I have a problem, when I try to do the VAR commands that you show here, stata gives out a "no observations r(2000);" error. What do you think can cause it?

[@7sanchito]

hi there, nice video. just in case do u have something about svar? i mean do file or an explanation about those models and their robustness tests we have to use? thank you.

[@yordanosbelete5566]

hi mike thank you so much. how do you merge the frede data with yahoo?

[@linhphamnguyenthuy3736]

Hey sir, I read about variance decomposition and unrestricted VAR model. Can u teach me the benefits of this and how to run in stata? 😭😭

[@m.abdurrahmanmehrabi4147]

I appreciate your efforts Dear Jonas, is it possible to apply the VAR model in panel data? I tried but the Stata gives me an error and says :(repeated time values in sample). can you help me with this case?

[@mikejonaseconometrics1886]

The command is "pvar" for panel VAR models. Here is an article that walks through the process: www.stata-journal.com/article.html?article=st0455 Good luck!

[@jackmurphy2403]

Hi Mike, this has been a huge help! The only problem I have encountered is that when trying to follow these steps with my data I get the error message 'repeated time values in sample' even though I have identified the panel variable and time variable using tsset, so that it is suitable for time series. Would you be able to shed any light on this?

[@mikejonaseconometrics1886]

Hi Jack, glad it was helpful. The standard var commands in Stata assume a pure time series, and do not operate on a panel. There are packages you can download for that purpose called pvar and xtvar. Use the command “ssc install pvar” to get the package and associated help files and documentation. Good luck!

[@jackmurphy2403]

@@mikejonaseconometrics1886 Thanks for the quick reply! Once installed is the process of running pvar largely the same as standard var i.e. can I follow the steps you've outlined in this video? Thanks again!

[@kadmielmartin-odoom6751]

Hello Mike. When i run my VAR model and tried to test for autocorrelation i got the error message "the lags of residuals may not be collinear with the dependent variables, or their lags." how do i fix this please

[@mikejonaseconometrics1886]

What command did you run to test for autocorrelation? I will try to replicate if you can let me know.

[@kadmielmartin-odoom6751]

@@mikejonaseconometrics1886 i had a variable that was highly correlated. Had a VIF of 15. When I took it out it run perfectly. Thank you very much

[@kadmielmartin-odoom6751]

@@mikejonaseconometrics1886 however I do have another question. How do you generate an Rsquared and Adjusted Rsquared measure for a VECM?

[@eleanorjonas9715]

hi dad!!!!

[@mikejonaseconometrics1886]

Hi Eleanor...