thank you always for your helpful videos. but in the last part after writing the system, an error message popped out stating that " C1 C2... are not defined" even though i did all the same exact steps.. what can i do regarding this matter?? thank u in advance
Thank you very much was this detailed and insightful video. I am doing a slightly more complex system in my panel data analysis and a syntax error pops up when I try to specify the lag of quadratic term in my regression. I was wondering if you could advise me on how to integrate this term into my system correctly.
Please after after estimating the 1st diff GMM I tried calculating the AR serial correlation but the system says the instrument was not saved I should reestimate using seecting KEEP WEIGHT. How do I do that please?
Thank you very much for your insightful video. I was wondering whether you know any paper that I can refer to to justify the method of choosing between first difference gmm and system gmm as you have demonstrated (comparing coefficient of the lagged dependent variable between 3 models)? Thank you very much once again.
I noticed that most authors who utilized system gmm in their studies hardly show that critical part as requisite analysis. However, I can share my preprint paper where I did the analysis to validate the choice of system gmm. Send a dm to my mail chekwus_vtus2008@yahoo.com or to my WhatsApp +2347069247145
Greeting Prof, your video has been my only guide to doing the system GMM in Eviews as there aren't any resources that explain the method. However, how do we get the system GMM equation sir? How are we supposed to generate an equation under the new object? Is it from the first difference dynamic panel estimator sir? Hoping for you reply which will be a greater guide for me to analyze my data with the system GMM. Thank you!
Thank you for your swift response to my enquiry. I am happy to inform you that after following your guidelines, I eventually succeeded in getting my System GMM output, though all the coefficients , namely C(1) C(2)............C(8) have turned out to statistically insignificant. My questions are as follows: What is the possible explanation for this “statistical mishaps”? Secondly, can these estimated coefficients be interpreted as short run or long run coefficients? Thirdly,, how can test for validity of the instruments and/or robustness of the estimated model? Thank you Prof. .
Please what is the remedy i used (-1) for the same number of independent variables but hitting okay the system says number of instrument greater than number of observations?
Greeting sir, your video was very informative, but i did not understand that how did you get the system GMM equation in the video time 10.23 onward, please guide. as i am following your video of System GMM but at this point i did not understand, how to make this equation, so that i can go to new object and make a system gmm box for equation? , sir, did you copy this equation from difference gmm ? your guidance will be appreciated.
What usually causes such error is too many lags when there are relatively small sample sizes. Reduce the number of lags and redo the estimation. Thanks
Thank you for the excellent display but please how you copied and pasted the last estimation you used to compute the sysGMM I have tried several times but I cant see the expression. That expression you copied and pasted and edited. thank you in advance for your respoense
Thank you for the video. How to frame an equation for system GMM. In the video you have copy pasted but I want to learn to frame an equation/estimations for system GMM. Can you help me? Or can you upload a video for the same.
Please what equation did you copy and pasted on the object screen. I mean the equation that you pasted and after you edited. Again which video did you interpret the result in?
how to interpret result in first difference gmm?if i get the same result for independent variable so means no problem for endogeneity right?how about if i get significant result for my lag variable?does it mean i have problem of endogeneity?
I have tried for system GMM using the equation as discussed. But I am getting error message showing that "D is an illegal or reserved name". Can you please suggest me what to do?
The dynamic panel wizard comes up immediately you change the estimation method to GMM. Which version of Eviews do you use? Student, University or Commercial?
Thank you Prof. Your videos are very interesting and educative. However, I have a problem with the last segment of your video(System GMM). Precisely, I created my GMM_SYM environment in my EViIEWS 10 as per your instructions but sadly, EVIEWS is giving an error message which might be related to how I specified my GMM System equations in the absence of a clear procedure on how to do so. Recall that your estimated system equations apparently imported into EVIEWS from somewhere; hence my inability to understand how you actually specified them. Grateful if you could guide me on how to correctly specify my system equations and/or overcome this “error message “ problem. Thank you once again.
Dear Salisu, to create your system GMM, add a system object just like I did in the video, then type in your system GMM equation in the system environment. I could have done mine by typing but I had to copy and paste mine to save time because I use the Limited Edition of recorder. Just take a cursory look at my system GMM equations and type yours in the system environment. Watch the video repeatedly and pause it where necessary to understand this clearly. Thanks.
Dear Sir, Is there any way to use System GMM model in time series data? In one of paper an anonymous reviewer has suggested to use System GMM model for time series data. However, as per as I know System GMM can be used only for Panel data when there is endogeneity problem. Could you suggest me what should I do in this case?
Yes! System GMM is applicable to time series. In fact, the example in this video is typical of time series conducted in a pool panel setup. The issue of endogeneity and simultaneity that underscores the use of Instrumental Variable (IV) regression is not limited to panel data models. Try to replicate the process in this video with your time series data and you’re good to go.
Although it all depends on your research problems, you may focus on the coefficients of the level variables represented by the first equation where the level of variables are estimated with the lagged differences as instruments. Thanks
Dear Max, thanks for commenting. If I understand your problem, you have variables that have negative values and log of negative values is a problem. If that's your problem, then all you need to do is to add a constant whose value is greater than or equal to the highest number with negative sign such that LogX=log(a+X), where X is the variable with negative values, a is a constant whose value is greater than or equal to the highest number with negative sign. For instance, if the highest number with negative sign in variable X is -4, then enter LogX=Log(4+X) in the Genr equation box. I hope this helps.
In Eviews language, c(1) means the first coefficient based on how the variables are ordered. Kindly look down below the results and see the equations where c(1), c(2), etc appear in the equations. If c(1) is the first coefficient of equation 1, then c(1) shows the average response of the dependent variable in equation 1 to the first independent variable in equation 1. Note that c(1) can also be a constant parameter depending on the equation.
Well done sir. I want to believe that the estimates obtained from this exercise (on System-GMM ) are short run. Please guide me on how to generate long run coefficients from system GMM using Eviews.
thank you always for your helpful videos. but in the last part after writing the system, an error message popped out stating that " C1 C2... are not defined" even though i did all the same exact steps.. what can i do regarding this matter?? thank u in advance
Thank you very much was this detailed and insightful video. I am doing a slightly more complex system in my panel data analysis and a syntax error pops up when I try to specify the lag of quadratic term in my regression. I was wondering if you could advise me on how to integrate this term into my system correctly.
Please after after estimating the 1st diff GMM I tried calculating the AR serial correlation but the system says the instrument was not saved I should reestimate using seecting KEEP WEIGHT. How do I do that please?
Thank you very much for your insightful video. I was wondering whether you know any paper that I can refer to to justify the method of choosing between first difference gmm and system gmm as you have demonstrated (comparing coefficient of the lagged dependent variable between 3 models)? Thank you very much once again.
I noticed that most authors who utilized system gmm in their studies hardly show that critical part as requisite analysis. However, I can share my preprint paper where I did the analysis to validate the choice of system gmm.
Send a dm to my mail chekwus_vtus2008@yahoo.com or to my WhatsApp +2347069247145
Greeting Prof, your video has been my only guide to doing the system GMM in Eviews as there aren't any resources that explain the method. However, how do we get the system GMM equation sir? How are we supposed to generate an equation under the new object? Is it from the first difference dynamic panel estimator sir?
Hoping for you reply which will be a greater guide for me to analyze my data with the system GMM. Thank you!
Thank you for your swift response to my enquiry. I am happy to inform you that after following your guidelines, I eventually succeeded in getting my System GMM output, though all the coefficients , namely C(1) C(2)............C(8) have turned out to statistically insignificant. My questions are as follows: What is the possible explanation for this “statistical mishaps”? Secondly, can these estimated coefficients be interpreted as short run or long run coefficients? Thirdly,, how can test for validity of the instruments and/or robustness of the estimated model? Thank you Prof.
.
Please what is the remedy i used (-1) for the same number of independent variables but hitting okay the system says number of instrument greater than number of observations?
Greeting sir, your video was very informative, but i did not understand that how did you get the system GMM equation in the video time 10.23 onward, please guide. as i am following your video of System GMM but at this point i did not understand, how to make this equation, so that i can go to new object and make a system gmm box for equation? , sir, did you copy this equation from difference gmm ? your guidance will be appreciated.
what are the differences between difference and orthogonal deviation in panel options?
Can you please guide me how did you wrote equation in system gmm thank you 😊
I wrote it already and copied it. I just pasted it so I can be faster.
How if u estimate sys gmm but error with notif "near singular matrix". How i can handle that?
What usually causes such error is too many lags when there are relatively small sample sizes. Reduce the number of lags and redo the estimation. Thanks
Thank you for this. Is there a way to select between one-step and two-step system GMM with this method?
For now, such option is not available in Eviews.
@@ChekwubeMadichie Thank you, how about the Arellano Bond test for autocorrelation?
Thank you for the excellent display but please how you copied and pasted the last estimation you used to compute the sysGMM I have tried several times but I cant see the expression. That expression you copied and pasted and edited. thank you in advance for your respoense
Thank you for the video. How to frame an equation for system GMM. In the video you have copy pasted but I want to learn to frame an equation/estimations for system GMM. Can you help me? Or can you upload a video for the same.
Facing Same Issue @ChekwubeMadichie please help us.
Thank you for the useful video. I have a question. How could we estimate panel VECM based on GMM in Eviews? Thank you so much!
Please what equation did you copy and pasted on the object screen. I mean the equation that you pasted and after you edited.
Again which video did you interpret the result in?
how to interpret result in first difference gmm?if i get the same result for independent variable so means no problem for endogeneity right?how about if i get significant result for my lag variable?does it mean i have problem of endogeneity?
I am from Nepal and I need additional more explanatory video on difference GMM and Sys GMM. How can I obtain?
Can we apply the technique for unbalanced panel?
do we need to include control variable if we run first difference gmm?
sir from where u copied the equation for system GMM.
How if i estimate with system gmm and error with "number of instruments grearger than number of observation"
Please watch the video carefully before you replicate. Reduce the number of instruments or increase the sample size.
I have tried for system GMM using the equation as discussed. But I am getting error message showing that "D is an illegal or reserved name". Can you please suggest me what to do?
Hi! I have multiple questions to ask about my sys-gmm program. is there any possibility to contact you in direct? thank you before.
chekwus_vtus2008@yahoo.com
Pls how do I bring out the dynamic panel data model wizard. It is not showing on my Eviews. Thanks.
The dynamic panel wizard comes up immediately you change the estimation method to GMM. Which version of Eviews do you use? Student, University or Commercial?
Thank you Prof. Your videos are very interesting and educative. However, I have a problem with the last segment of your video(System GMM). Precisely, I created my GMM_SYM environment in my EViIEWS 10 as per your instructions but sadly, EVIEWS is giving an error message which might be related to how I specified my GMM System equations in the absence of a clear procedure on how to do so. Recall that your estimated system equations apparently imported into EVIEWS from somewhere; hence my inability to understand how you actually specified them. Grateful if you could guide me on how to correctly specify my system equations and/or overcome this “error message “ problem. Thank you once again.
Dear Salisu, to create your system GMM, add a system object just like I did in the video, then type in your system GMM equation in the system environment. I could have done mine by typing but I had to copy and paste mine to save time because I use the Limited Edition of recorder. Just take a cursory look at my system GMM equations and type yours in the system environment. Watch the video repeatedly and pause it where necessary to understand this clearly. Thanks.
Dear Sir, Is there any way to use System GMM model in time series data? In one of paper an anonymous reviewer has suggested to use System GMM model for time series data. However, as per as I know System GMM can be used only for Panel data when there is endogeneity problem. Could you suggest me what should I do in this case?
Yes! System GMM is applicable to time series. In fact, the example in this video is typical of time series conducted in a pool panel setup. The issue of endogeneity and simultaneity that underscores the use of Instrumental Variable (IV) regression is not limited to panel data models. Try to replicate the process in this video with your time series data and you’re good to go.
Thank you so much this is so useful, can you share with me the data??
Kindly join my community and become a member to enjoy certain privileges and support in your research
Mine is showing me “near singular matrix”
are lfind and lfaid independent variables?
Yes
share the workfile
and how to do the arellano and bond test?
This is based on the assumption that none of such problems exists.
Will our results be reliable if we assume so? Rather I did arellano and bond test when I applied difference GMM in Eviews. @@ChekwubeMadichie
why there isnt name of variables . sir i wish u have interpretate the result as well
I will consider interpretation in my next video.
@@ChekwubeMadichie thank you
how do we analyse these results given that we have two coefficients for each variable?
Although it all depends on your research problems, you may focus on the coefficients of the level variables represented by the first equation where the level of variables are estimated with the lagged differences as instruments. Thanks
Sir i need your small help, my model is ready but negative log of value is problem for me.
Dear Max, thanks for commenting. If I understand your problem, you have variables that have negative values and log of negative values is a problem. If that's your problem, then all you need to do is to add a constant whose value is greater than or equal to the highest number with negative sign such that LogX=log(a+X), where X is the variable with negative values, a is a constant whose value is greater than or equal to the highest number with negative sign. For instance, if the highest number with negative sign in variable X is -4, then enter LogX=Log(4+X) in the Genr equation box. I hope this helps.
@@ChekwubeMadichie Thanks for your reply. I will try to fix data.
Thank you for your video, How I can interpret the results in system GMM output? what c(1).... means?
In Eviews language, c(1) means the first coefficient based on how the variables are ordered. Kindly look down below the results and see the equations where c(1), c(2), etc appear in the equations. If c(1) is the first coefficient of equation 1, then c(1) shows the average response of the dependent variable in equation 1 to the first independent variable in equation 1. Note that c(1) can also be a constant parameter depending on the equation.
Well done sir. I want to believe that the estimates obtained from this exercise (on System-GMM ) are short run. Please guide me on how to generate long run coefficients from system GMM using Eviews.
When I estimate GMM Time series HAC, why my Sintax error?
Make sure you enter all the symbols correctly in the model specification. Watch the video carefully and repeatedly to understand better.
Thank you @@ChekwubeMadichie
Welcome