I’ve been doing time series analysis for more than a semester and was so confused before I found this video. Thank you so much! Just enough information to understand but not too much that you get lost
Thanks for your positive feedback! I am glad it helped you. Feel free to check my website www.jdeconomics.com where you will find all the available material and topics so far. Good luck with your studies/model. Regards, JD
Excellent. Please 2 questions. Is it desirable that there be Granger Causality to run a VAR? And do you have any routine for running an SVAR with the Beaudry and Portier (2006) method? Thanks.
Hello! Thanks for your message. What type of content would you like on SAS? I have a lot of tutorials to make and still have to get into Dynare with DSGE models but I will try to add it on the list. Feel free to subscribe to my channel (if you haven't) so you stay tuned! Kind Regards, JD Econ.
Hello! If our residuals are autocorrelated, what next steps could we take to fix them? (Also thank you so much for the informative presentation. It helps a lot!)
Hello Everyone! Thanks for watching! ✅ You can Buy the Stata DO File + Slides at: jdeconomicstore.com/b/var-model-in-stata 📈Dataset to download [Free] : (Unemployment and Fed Rate - USA): jdeconomicstore.com/b/var-model-in-stata 🌐Here is the link for Part 2 of the tutorial: ua-cam.com/video/plZREZ3i5Tc/v-deo.html 🌐The tutorial is also available in EViews at: ua-cam.com/video/SbE8ns0oOTs/v-deo.html 🌐Please subscribe to my channel for more tutorials! ua-cam.com/channels/5P21WGFO4WRUlAiGLcwymg.html Thanks a lot! JDEcon.
Is it VAR can measure volatility? I have my thesis " titled: consumption volatility oil and energy." We used the model VAR for measuring volatility but I'm confused if it suitable model for volatility. ? Hoping for feedback and recommendations what model we use? Thanks you. It's VAR or Garch?. it's very helpful to us👍.
Thank you for this tutorial. Very insightful and easy to understand. I noticed you estimated the VAR model using the variables at difference since they are not stationary at levels. Does this mean that a VAR model can not be estimated using non stationary variables?
Depends. ARDL allows for different lags, while VEC is used for cointegrated variables. VAR is for models where all the dependent variables have lags of the same order. I hope that helps. Regards
High Quality Sound and Editting, a very good video on the entry-level econometric model. Looking forward for more basic and advance models. Been struck on low quality video for as long as remembered, please keep the video coming on different social sci econometric models like panel models, NARDL, porbit, etc.
Great video ! it helped me a lot! I was wondering why most of the researchs on monetary policy impacts don't use first differences in their structural var models?
Hi! Thanks for your message. I can’t speak for them. I am sure that they clarify the reasons behind their model on the methodology section of their paper. Some clarify that they are looking for long run relationships and not for short run dynamics. Regards, JD.
@@JDEconomics thanks! yeah in the research that I read it wasn’t said tho but I saw that if you first difference them you lose important information. Thats it its for long term Impulse responses i finally have the word for it.
Thank you so much for this video, i have a question, what is one supposed to ddo when the VAR is not stable, because that is the point where i am really stuck
Hello Juan! Thank you for your great video! I found it extremely helpful for conducting my thesis! I have one question though. If one of my variables is non stationary and the rest are stationary, do i need to take the first differences for all the variables to be able to conduct VAR, or do i just take the 1st differences for the one variable and conduct VAR with the rest of the variables as they are and the 1st differences of the one variable? Thank you in advance.
Very useful video Juan! After doing first differences the ADF test tells me my variables are stationary. The AIC, HQ and BIC tell me to use 2 lags. But when I do the residual diagnosis after estimating the VAR (2) model it tells me my VAR model has autocorrelation for 1, 2, 3,4.... all lags. How can I solve this problem?
Hello Luis, thanks for your positive feedback. I can think of two outcomes here. One is trying with the amount of lags corresponding to the frequency of your data (i.e., if quarterly, use 4 lags). Perform the test again then, and check. Alternatively, there are times when the residuals show a big spike at a certain point. (i.e., 2008 crisis). Sometimes including a dummy variable (year of the spike = 1) as an exogenous variable in the model can help us mitigate your problem. Kind Regards, JD.
Thanks for the insightful video. Although can you clarify one question,? Im running a VAR model with 30 years data and 5 variables. But when i check for autocorrelation, Stata shows "the exogenous variables may not be collinear with the dependent variables, or their lags". What to do now?
Hello Thinh! Thanks for your feedback! I will look into adding notes! If you active subtitles ( CC on), youtube will generate the subtitles for you. I will keep submitting tutorials, so feel free to subscribe (if you haven’t), so you get notified! Thanks! JD
Hey, typically you have to specify the series is in differences. So in your paper, you can either use the symbol △Yt or you can write " The variable Y is in first differences". I should have probably written the model using △Unemployment=............ However, since I was orally explaining that the variable was in differences, I think it was clear. But definitely, make sure you use △ or write it in the "model section" of your paper. Thanks for the observation! Regards, JD
Is it VAR can measure volatility? I have my thesis " titled: consumption volatility oil and energy." We used the model VAR for measuring volatility but I'm confused if it suitable model for volatility. ? Hoping for feedback and recommendations what model we use? Thanks you. It's VAR or Garch?
Vecm models are for variables that are non stationary in levels but are cointegrated. Means they have a long run relationship that results in a stationary residual series. I will do a tutorial in some time. Regards, JD
Beautiful explanation. only one quiry. when we can apply VAR or Panel Var model? is it when all variables are I(1) but not co-integrated ? since if there is co-integration among the variables then we could apply VEC Error Correction Model. another thing if the variables are integrated in different order say I(0) and I(1) we could apply ARDL model. please clarify my doubt and also requesting to prepare one video for Panel VAR model.please response.
Thanks Ivana for your positive comment! You can do so. I think I haven't covered it in the video. However, note that heteroskedasticity won't affect the consistency of the var coefficient estimates. You can test for it. I hope that helped to clarify your question. Kind Regards, JD.
Excellent video, but without testing Johansen's cointegration test, how can you straightforwardly go for VAR estimation? If the variables are cointegrated you cannot estimate VAR model.
Hi. Thanks for your feedback! It’s just an illustrative video to portrait var models. It is a good practice to check for cointegration first. I encourage everyone to do so. Just for the sake of time, I didn’t cover cointegration in the video. Cheers,
Hey. I haven’t tried it before I think. I quickly checked the manual, and it doesn’t seem to include post-estimation statistics produced by Stata. You can check for autocorrelation and the normality of errors, but there are no directions about heteroscedasticity (see manual: www.stata.com/manuals/tsvarpostestimation.pdf). You may try the command ‘estat hettest’ and see if that works. This command is normally used after regression estimation, and I don’t think it would work after a VAR model. Regards,
Thanks so much for this video. How can I introduce or arrange multiple-year data? (for example year ESG data and Financial Performance of 50 companies)
I am glad to hear you liked the video! And I'm sorry, I am not sure I understand your question.
2 роки тому
@@JDEconomics thanks for getting back to me. I want to evaluate the causality correlation between two variables that are changing during the time for 50 companies. But my question is how can arrange the data? I will have 50 point in MC at 2018, 50 point at 2019.. etc
Thank you! Is it VAR can measure volatility? I have my thesis " titled: consumption volatility oil and energy." We used the model VAR for measuring volatility but I'm confused if it suitable model for volatility. ? Hoping for feedback and recommendations what model we use? Thanks you. It's VAR or Garch?
Great video :) I am using you video as a template for my thesis but I have an exogenous variable. When I perform the Granger test it excludes all my variables. Any idea? :/
Thank you for for this video. I would love to ask the following. If I have a number of variable say 6. can i test for causality for all at once in a model or just a pair at a time. Lastly, I am interested is showing the coefficient, e.g Y causes X with what percentage. How do i read the coefficient or precisely normally which values are recorded in the table.
Hey, with the causality test you can only determine if they cause each other (not how much). However, you can take a look at the variance decomposition to see how much of the variation in each variable is due to each of the endogenous variables in the system. Regards
Hey, thank you for this video. I am facing an issue of "no observations" when I use "dfuller unemployment". I used tsset and set it to quarterly and have tried many other tricks mentioned in the statalist website but could not resolve it. can you please help me?
Hiya, thank you so much for the video! I have a question: for the output of the "var dunemp drate, lags(1/2)" command in the "how to estimate the var" section, you said that dunemp is the dependent variable. But in the "our example" section, I thought the unemployment rate is the independent variable since it is what leads to change in the fed rate. So I'd like to know if the independent or dependent variable comes after the "var" command.
What if in the varlmar/Lagrange-multiplier test in lags 2 there is an autocorrelation, what else do we need to do in order to fix or at least say that there is no autocorrelation at lag order
hello sir, i would like to ask about the VAR estimate. What should i do if the that we choose is 4 and when i estimate with "varsoc" the best estimate using lag 4, and i try again with lag 5 and the best estimate changes using lag 5. Then, what should i choose?
thanks for the video, it was very informative. I am currently facing an issue where when entering “varlmar” instead of getting the table output I get an error message saying “the lags of residuals may not be collinear with the dependent variables, or their lags” do you have any idea how to solve this ?
Hey, I am not sure about that one. Have you checked Stata manual? Chances are the amount of lags used in the var code are causing that issue. Maybe use different number of lags. Other than that, I couldn’t tell you what’s the solution. Regards, JD
I would use 4 lags if its quarterly data, 1/2 lags if it is monthly data, and so on. At the end of the day, you need to ensure there is no autocorrelation at the specified lag and that the impulse response functions look smooth and sensible. Regards, JD.
Not very common to see if the data is stationary. It can happen that there are some critical years (70s, early 90,2000,2008, 2020) you can try including a dummy if you see a spike. Regards
@@JDEconomics Thanks for your reply. I also have another query about interpreting var model when the prob>chi2 is >.05. If I run a simple var model: var x y , the prob>chi2 is >.05 i.e. non-significant, does this mean that I cannot use this model?
i did that but the problem is varsoc has given me options for lag length which a 3 and 4 but when i use those lags the stability condition is not satisfied but when I use 1 and 2 which is not an option from the varsoc report, the stability condition is then satisfied 😭
I have a problem when doing LM autocorrelation test. Whenever I write a command "varlmar" ,Stata says that the exogenous variables may not be collinear with the dependent variables, or their lags. What does that mean? Can you help me with this error?
@@malihawasim the choleski decomposition is a decomposition used to identify the contemporaneous causal effects between the variables of a var model. The default identification restriction used by Stata shows the short run causal relationships. It’s a lower triangular matrix. Hope that helps. Regards!
Hey, very useful video although I really need your help for my quantitative techniques project, can you help me please I need to apply var on some macroeconomic variables, I found stata to be user friendly but I'll have to pay for it can you help me with the project?
I’ve been doing time series analysis for more than a semester and was so confused before I found this video. Thank you so much! Just enough information to understand but not too much that you get lost
You're very welcome!
Never find such a useful video on UA-cam. Thank You Sir keep it up.
Thanks. Please share it with your network and feel free to subscribe to the channel.
Regards.
Thank you so much for such a condense and informative lecture! Your concrete examples really help a lot!
Thanks for your positive feedback! I am glad it helped you. Feel free to check my website www.jdeconomics.com where you will find all the available material and topics so far. Good luck with your studies/model. Regards, JD
Very comprehensive and well documented JD. Thanks!
Thanks! Please feel free to share my channel with your friends/social media. Regards, Jd
Thank you so much for your video, very helpful and informative
Glad it was helpful!
Excellent. Please 2 questions. Is it desirable that there be Granger Causality to run a VAR? And do you have any routine for running an SVAR with the Beaudry and Portier (2006) method? Thanks.
very appreciative!!
please make some series on SAS
Hello! Thanks for your message. What type of content would you like on SAS? I have a lot of tutorials to make and still have to get into Dynare with DSGE models but I will try to add it on the list. Feel free to subscribe to my channel (if you haven't) so you stay tuned!
Kind Regards,
JD Econ.
@@JDEconomics
some major regression models (familiar with master programme)
thank you so much for these incredible videos, you save my life😭😭🥰🥰
Great to hear! Please feel free to subscribe to the channel and share it to your friends! That helps me a lot 😀
Good luck!
Invaluable video! Thanks!
Thanks! Please share it with your network. Many thanks!
Hello! If our residuals are autocorrelated, what next steps could we take to fix them? (Also thank you so much for the informative presentation. It helps a lot!)
great lecture Juan thank you for sharing your material keep up the great work xxx
Thanks!
Thanx for your videos and explanations. Is it applicable to use VAR in case of I(0) variables?
Yes! Absolutely.
Very well explained.
Thanks a lot! Please feel free to subscribe to my channel and share it with your friends/social network. I wish you good luck in your studies/work!
Hello Everyone! Thanks for watching!
✅ You can Buy the Stata DO File + Slides at:
jdeconomicstore.com/b/var-model-in-stata
📈Dataset to download [Free] : (Unemployment and Fed Rate - USA):
jdeconomicstore.com/b/var-model-in-stata
🌐Here is the link for Part 2 of the tutorial: ua-cam.com/video/plZREZ3i5Tc/v-deo.html
🌐The tutorial is also available in EViews at: ua-cam.com/video/SbE8ns0oOTs/v-deo.html
🌐Please subscribe to my channel for more tutorials!
ua-cam.com/channels/5P21WGFO4WRUlAiGLcwymg.html
Thanks a lot!
JDEcon.
Is it VAR can measure volatility? I have my thesis " titled: consumption volatility oil and energy." We used the model VAR for measuring volatility but I'm confused if it suitable model for volatility. ? Hoping for feedback and recommendations what model we use? Thanks you. It's VAR or Garch?. it's very helpful to us👍.
good
Thank you for this tutorial. Very insightful and easy to understand. I noticed you estimated the VAR model using the variables at difference since they are not stationary at levels. Does this mean that a VAR model can not be estimated using non stationary variables?
Correct
Thank you for your video. I have a question. I read studies used firstly ARDL then used VAR or VECM. When do we use firstly ARDl then VAR orVECM?
Depends. ARDL allows for different lags, while VEC is used for cointegrated variables. VAR is for models where all the dependent variables have lags of the same order. I hope that helps. Regards
Thank you for your response.
thank you for this very informative video! i was wondering if the process is the same when we have more than two independent variables?
Yes! All the same
High Quality Sound and Editting, a very good video on the entry-level econometric model. Looking forward for more basic and advance models. Been struck on low quality video for as long as remembered, please keep the video coming on different social sci econometric models like panel models, NARDL, porbit, etc.
Thanks for the feedback! Regards, JD
Thank you so much, it was helpful
You're welcome!
Great video ! it helped me a lot! I was wondering why most of the researchs on monetary policy impacts don't use first differences in their structural var models?
Hi! Thanks for your message. I can’t speak for them. I am sure that they clarify the reasons behind their model on the methodology section of their paper. Some clarify that they are looking for long run relationships and not for short run dynamics. Regards, JD.
@@JDEconomics thanks! yeah in the research that I read it wasn’t said tho but I saw that if you first difference them you lose important information. Thats it its for long term Impulse responses i finally have the word for it.
One reason may be the variables are stationary at level that is all are I(0) variable.
Thanx for your videos. Why number of observations reduced when performing the VAR model?
your explain is very cleary and interested. how can I find your ep. of SUR model in stata?
I don;t have such video. Sorry about that!
Thank you (VERY) much, waiting
for SVAR
Thanks
Thank you so much for this video, i have a question, what is one supposed to ddo when the VAR is not stable, because that is the point where i am really stuck
Variables may not be stationary. Check that out
Very lucid presentation
Thanks!
Hello Juan! Thank you for your great video! I found it extremely helpful for conducting my thesis!
I have one question though. If one of my variables is non stationary and the rest are stationary, do i need to take the first differences for all the variables to be able to conduct VAR, or do i just take the 1st differences for the one variable and conduct VAR with the rest of the variables as they are and the 1st differences of the one variable?
Thank you in advance.
Yes, absolutely! Just in the non stationary. Good luck!
Excellent video, can you do a video on ARDL, GDP and Capital Investment Budget.
Hello, thanks for your message. I can add it to my list of videos. Feel free to subscribe to the channel so you get notified! Thanks
Thanks for the very informative video. How can I conduct VAR with panel data also?
Thanks! I don’t have a tutorial on that yet. Hopefully in the near future! Regards
Very useful video Juan!
After doing first differences the ADF test tells me my variables are stationary. The AIC, HQ and BIC tell me to use 2 lags. But when I do the residual diagnosis after estimating the VAR (2) model it tells me my VAR model has autocorrelation for 1, 2, 3,4.... all lags. How can I solve this problem?
Hello Luis, thanks for your positive feedback. I can think of two outcomes here. One is trying with the amount of lags corresponding to the frequency of your data (i.e., if quarterly, use 4 lags). Perform the test again then, and check. Alternatively, there are times when the residuals show a big spike at a certain point. (i.e., 2008 crisis). Sometimes including a dummy variable (year of the spike = 1) as an exogenous variable in the model can help us mitigate your problem.
Kind Regards,
JD.
What a life saver, thank you so much! :)
Hi! thanks for your positive feedback. I am glad it helped you. Best Regards, JDEconomics
Hi! Very good video. I have a question, what happens when de "varsoc" test indicates just one lag? Thanks
You can use one lag! Use The option: lags(1)
Thanks for the insightful video. Although can you clarify one question,? Im running a VAR model with 30 years data and 5 variables. But when i check for autocorrelation, Stata shows "the exogenous variables may not be collinear with the dependent variables, or their lags". What to do now?
Thanks for your videos. In addition, could you add notes to videos? My listening skill is quite bad. I look forward to your new videos.
Hello Thinh! Thanks for your feedback! I will look into adding notes! If you active subtitles ( CC on), youtube will generate the subtitles for you. I will keep submitting tutorials, so feel free to subscribe (if you haven’t), so you get notified! Thanks! JD
Thank you for the great video
Thanks Yassine for your positive feedback! Good luck! JD
Thank you so much! I have one question about the specification of the VAR model. Shall we specify the model in levels? given the series are I(1)
Hey, typically you have to specify the series is in differences. So in your paper, you can either use the symbol △Yt or you can write " The variable Y is in first differences". I should have probably written the model using △Unemployment=............ However, since I was orally explaining that the variable was in differences, I think it was clear. But definitely, make sure you use △ or write it in the "model section" of your paper. Thanks for the observation! Regards, JD
Dear Juan
Can you please make a estimation video of Bayesian VAR models in STATA , it would be really helpful
Hi! I can’t promise anything for the short time since I am busy finishing the DSGE course. I hope to start more series shortly. Regards,
Is it VAR can measure volatility? I have my thesis " titled: consumption volatility oil and energy." We used the model VAR for measuring volatility but I'm confused if it suitable model for volatility. ? Hoping for feedback and recommendations what model we use? Thanks you. It's VAR or Garch?
Garch models are normally used for volatility! I have some tutorials about ARCH and GARCH Models. Cheers!
May I have a question?
What is the different between cointegrated VAR (CVAR) and VAR model? And then, how can run CVAR on Stata.
Thank you.
Vecm models are for variables that are non stationary in levels but are cointegrated. Means they have a long run relationship that results in a stationary residual series. I will do a tutorial in some time. Regards, JD
Beautiful explanation. only one quiry. when we can apply VAR or Panel Var model? is it when all variables are I(1) but not co-integrated ? since if there is co-integration among the variables then we could apply VEC Error Correction Model. another thing if the variables are integrated in different order say I(0) and I(1) we could apply ARDL model. please clarify my doubt and also requesting to prepare one video for Panel VAR model.please response.
Thank you so much sir.
i had learned a lot from this vedio.
sir how can we export our results from stata 17 to world file?
Highlight, copy-paste.
Fantastic! Thank you so much!!
Great! Thanks! Cheers, Jd
Thank you for the great video! Shouldn't we also use formal tests for heteroscedasticity?
Thanks Ivana for your positive comment! You can do so. I think I haven't covered it in the video. However, note that heteroskedasticity won't affect the consistency of the var coefficient estimates. You can test for it.
I hope that helped to clarify your question.
Kind Regards,
JD.
Excellent video, but without testing Johansen's cointegration test, how can you straightforwardly go for VAR estimation? If the variables are cointegrated you cannot estimate VAR model.
Hi. Thanks for your feedback! It’s just an illustrative video to portrait var models. It is a good practice to check for cointegration first. I encourage everyone to do so. Just for the sake of time, I didn’t cover cointegration in the video. Cheers,
Hi, very good video, but i have 1 question, its about heteroscedasticity, how can I found the heteroscedasticity in a VAR model, hope you can answer.
Hey. I haven’t tried it before I think. I quickly checked the manual, and it doesn’t seem to include post-estimation statistics produced by Stata. You can check for autocorrelation and the normality of errors, but there are no directions about heteroscedasticity (see manual: www.stata.com/manuals/tsvarpostestimation.pdf). You may try the command ‘estat hettest’ and see if that works. This command is normally used after regression estimation, and I don’t think it would work after a VAR model. Regards,
Thanks for the video. For the stability test for the VECM, what does it mean if Stata says "The VECM specification imposes 4 unit moduli"? Thank you!
Very useful video
Thanks Jhabindra! I will be submitting new video, feel free to subscribe (if you haven’t), so you get notified. Regards, JD!
EXCELLENT
Many thanks!
Thanks so much for this video. How can I introduce or arrange multiple-year data? (for example year ESG data and Financial Performance of 50 companies)
I am glad to hear you liked the video! And I'm sorry, I am not sure I understand your question.
@@JDEconomics thanks for getting back to me. I want to evaluate the causality correlation between two variables that are changing during the time for 50 companies. But my question is how can arrange the data? I will have 50 point in MC at 2018, 50 point at 2019.. etc
Great job sir
Thanks!
Thank you! Is it VAR can measure volatility? I have my thesis " titled: consumption volatility oil and energy." We used the model VAR for measuring volatility but I'm confused if it suitable model for volatility. ? Hoping for feedback and recommendations what model we use? Thanks you. It's VAR or Garch?
Great video :) I am using you video as a template for my thesis but I have an exogenous variable. When I perform the Granger test it excludes all my variables. Any idea? :/
Thank you for for this video. I would love to ask the following.
If I have a number of variable say 6. can i test for causality for all at once in a model or just a pair at a time.
Lastly, I am interested is showing the coefficient, e.g Y causes X with what percentage. How do i read the coefficient or precisely normally which values are recorded in the table.
Hey, with the causality test you can only determine if they cause each other (not how much). However, you can take a look at the variance decomposition to see how much of the variation in each variable is due to each of the endogenous variables in the system. Regards
Du bist toll.
Thanks!
GREAT VIDEO !!!!
Thanks! Feel free to share it with your close ones! Best Regards, JD
Hey, thank you for this video. I am facing an issue of "no observations" when I use "dfuller unemployment". I used tsset and set it to quarterly and have tried many other tricks mentioned in the statalist website but could not resolve it. can you please help me?
Thanks. I don’t know as I don’t know what’s your dataset.
The same dataset that you are using, i downloaded it from your website.
Nevermind, i resolved it! Thanks@@JDEconomics
Hiya, thank you so much for the video! I have a question: for the output of the "var dunemp drate, lags(1/2)" command in the "how to estimate the var" section, you said that dunemp is the dependent variable. But in the "our example" section, I thought the unemployment rate is the independent variable since it is what leads to change in the fed rate. So I'd like to know if the independent or dependent variable comes after the "var" command.
Hi, all variables are endogenous in var models. The reason I put the variables in that order is due to the identification strategy. Regards
Thank you so much.
You are very welcome! You can buy the material in case you are interested! Good luck with your studies! JD
Very helpful
Thanks!
the best !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Thanks a lot! Please share it with your network 😃
thank you very much
No worries!
That for video! Is it applicable for 9 explanatory variables?
You can, but as you can imagine, ordering 9 variables is a bit more complicated than 3/4 variables. Good luck! JD.
hello sir, what if the data use lag 5 but when we write varlmar only 2 lag are appears?
Why don’t you transform them to log form before first differencing?
What if in the varlmar/Lagrange-multiplier test in lags 2 there is an autocorrelation, what else do we need to do in order to fix or at least say that there is no autocorrelation at lag order
in VAR model it is important to generate the IRFs, could please highlight how can we obtain the IRFs?
Thanks
hello sir, i would like to ask about the VAR estimate. What should i do if the that we choose is 4 and when i estimate with "varsoc" the best estimate using lag 4, and i try again with lag 5 and the best estimate changes using lag 5. Then, what should i choose?
Hi! Use normally what the lag length suggests. Also, take into account that adding more lags may affect the impulse response functions smoothness.
thanks for the video, it was very informative. I am currently facing an issue where when entering “varlmar” instead of getting the table output I get an error message saying “the lags of residuals may not be collinear with the dependent variables, or their lags” do you have any idea how to solve this ?
Hey, I am not sure about that one. Have you checked Stata manual? Chances are the amount of lags used in the var code are causing that issue. Maybe use different number of lags. Other than that, I couldn’t tell you what’s the solution. Regards, JD
What do you suggest if all the criterion suggest 0 lags?
I would use 4 lags if its quarterly data, 1/2 lags if it is monthly data, and so on. At the end of the day, you need to ensure there is no autocorrelation at the specified lag and that the impulse response functions look smooth and sensible. Regards, JD.
Hello!
What if there is autocorrelation? How to manage with it?
Not very common to see if the data is stationary. It can happen that there are some critical years (70s, early 90,2000,2008, 2020) you can try including a dummy if you see a spike. Regards
can you make panel var with stata please
If some of my variables are stationary but some are not, should I change all of the to differences or just the ones that are non-stationary?
Change the non stationary ones. Regards.
Is the unemployment shock here an endogenous variable or exogenous? I want to include an exogenous shock to my model (covid) would that be okay?
Unemployment is an endogenous variabke
@@JDEconomics Thank you. Excellent video as always
@@sarahm9140 thanks!
HI,what's the difference bewteen
var dunemp drate,lags(1/2) and
var unemp rate,lags(1/2)
can I just run
var unemp rate,lags(1/2)
Thank you
Unemployment is just the raw variable, dunemp is the variable after applying the necessary transformations to make it stationary. Regards
@@JDEconomics Thanks for your reply. I also have another query about interpreting var model when the prob>chi2 is >.05.
If I run a simple var model: var x y , the prob>chi2 is >.05 i.e. non-significant, does this mean that I cannot use this model?
For panel data do as the same or not ?
for the stability test what do i do if the stability condition is not satisfied
Try a model with different lags
i did that but the problem is varsoc has given me options for lag length which a 3 and 4 but when i use those lags the stability condition is not satisfied but when I use 1 and 2 which is not an option from the varsoc report, the stability condition is then satisfied 😭
I have a problem when doing LM autocorrelation test. Whenever I write a command "varlmar" ,Stata says that the exogenous variables may not be collinear with the dependent variables, or their lags. What does that mean? Can you help me with this error?
Hey! I’m not sure about that error. Maybe you have few lags, or not many observations. I can’t think of any other problem! Sorry, JD
@@JDEconomics thank you for your reply, the problem is solved. Now I want to know SVAR and cholesky Var similar?
@@malihawasim the choleski decomposition is a decomposition used to identify the contemporaneous causal effects between the variables of a var model. The default identification restriction used by Stata shows the short run causal relationships. It’s a lower triangular matrix. Hope that helps. Regards!
@@JDEconomics Thanks a bunch.
Hey, very useful video although I really need your help for my quantitative techniques project, can you help me please I need to apply var on some macroeconomic variables, I found stata to be user friendly but I'll have to pay for it can you help me with the project?
You can send me an email! Regards
I love you brother
Thanks!
how can i use SVAR in stata please