TH-cam 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, help me stay online.
Thank you so much teacher for this videos . these videos really helped me . Thank you again for the attention that you pay on my comment and my question
Thank you very much. I wonder if you show us how to conduct Panel sub-sample or sub-region analysis in Eviews starting from spliting sample or creating region id. You may use Sub-Saharan Africa as case for this. Thank you in advance!
Thank you Dr for this analysis. I have a question, the issue is I am using ARDL technique which is involving a dummy variable, everything is fine even the F statistic and the diagnostic tests BUT the problem is on cusum and cusum of squares test, it is not showing the two red lines and the blue line BUT rather it is showing a bar graph. IF i remove the dummy, the two red lines and a blue line on the cusum and cusum of squares will be shown but not when i use dummy. Also, is it ok if a person uses Ramsey test for stability when using ARDL instead of cusum and cusum of squares? Your response is appreciated.
Hi Wan, watch my videos on DUMMY VARIABLES, Error Component Models and Panel Sub-Samples. I explained the coefficient of dummy variables in those clips and I'm sure I explained in this video too. Thanks
Thank you for this tutorial video. I would like to ask a question. Why do you need to interact the variables with the time dummy? Do you have any book and or journal references on this? Thank you.
Hi Michael, I don't have any reference but you can always search online for textbooks or articles that modelled structural breaks. I only showed how it's done.
thanks prof, amazing video. but i have a question, when i try to testing unit root for variabel dummy (where, 0 before covid and 1 after covid), the result is "singular matrix". can you help me and what should i do? thanks before
Hello, first of all I would like to thank you for your videos as they were very helpful for my studies and research. I would like to ask if there is a problem when one wants to incorporate a break that occurs in his last observations, meaning that there are not many observations that were affected. Can many "zeros" in the dummy variables (an example would be 95 out of 100 observations take the value zero) be a problem in a model (ECM, ARDL and so on)? Moreover, would it be logical if one tried to also examine how the lagged values of the structural break could affect a model (instead of the dummy variable itself) or there is not a sufficient interpretation of this?
Hello CrunchEconometrix, thanks for this amazing tutorial. My question is, if I run my bound test and the results shows no long run relationship between my variables, do i need to use dummy variables for my analysis if thier is still the presence of structural break? Lastly what are the adequacy test or diagnostic checks that I need to subject my Short run ARDL model to in order to know whether it is good for interpretation. Thank you.
Thanks Berk for the positive feedback...appreciated! Your queries are a bit confusing. Several unrelated approaches are muddled up. Please be more explicit. Using dummy variables may be specific to your study. I also covered diagnostics (hetero, serial corr, normality, stability) in my ARDL videos. Kindly watch them again.
Hello Sir. I have a question and if you can answer, I will be very happy for that. I have got two dummy variables in my ARDL model. When I put my two dummy variables in the box of dependent variable followed by list of dynamic regressors, it works. However, when I put my dummy variables in the list of fixed regressors, Eviews gives a warning "near singular matrix error". What should I do ? Thank you for your videos.
Hi professor, thank u so much for your video it is very helpful. i am running an ARDL model with dummy variable for the break point specified by the chow and quandlt test for breaks. However I am wondering why the dummy coefficient is insignificant although the tests proved the presence of break?
Hi Ingy, it is possible for the break dummy coefficient to be statistically not significant. It implies that it has no impact on influencing the behaviour of the dependent variable.
Thank you I appreciate that , I ask opinion from others and I research for my own and I found that dummy variable is stationary by nature in first difference, are this is correct ?
hi professor, to assess the importance of each independent variable in determining the dependent should I use the standardized coefficients? or how to quantify each variable (or group of variables) contribution ?
Thank you for sharing these videos, i study a lot since follow your channel. May i ask you some questions? 1. can I use the dummy variables withe panel ARDL and panel VAR model?
Hi Xu, thanks for the compliments. Can't say exactly cos I've never tried it. But do further online findings on it, give it an attempt and see what you come up with?
Hi Kenechukwu, due to abuse I disabled the link. Some datasets are now available on my website upon payment of a token. Here is the link cruncheconometrix.com.ng/shop
How to OLS regression this ARDL model? While Eviews is making a parsimonious specification of it, I thought it would be better to regress like you did in the OLS approach video
Hi Dr Ngozi, I have a question related to control variable. My supervisor asked me to use inflation in my model as a control variable so how do I incorporate the control variable in Eviews. Thanks
@@CrunchEconometrix thanks Dr Ngozi for your support, one final question on this, in Eviews help on control variable, they said to put !, So if I understood correctly, in the estimation equation, I put my dependent and independent variable and ! inflation because this is my control variable
Dear professor, thank you very much for another great video. I have a question. When we write a paper, should we interpret coefficients of dummy variables as a short-run or long-run relationship?
Hi Djordje, interpretation will depend if at the end of the estimation the coefficient appears in the long-run and/or short-run parts of the estimation output.
Guys, how did we get dum_lndc from dum? Then if we have say six independent variables, will we have dum interaction with the six independent variables? Thank you in advance.
@@CrunchEconometrix Thank you. I finally got the explanation. Lastly the dummy variable interaction can it be extended to 7 independent variables or they are too many?
@@CrunchEconometrix My study looks at how mobile money has impacted on private investment. The independent variables are GDP= the percentage change in real GDP; GI = the ratio of public-sector investment to GDP; CR = the change in credit to the private sector; CPI = the percentage change in consumer price index; TOT = the change in terms of trade; RER = the index of real exchange rate; OPEN = Openness of the economy expressed as exports plus imports over GDP FI = Value of MPESA transactions Elec_d = Election year dummy e When I follow your procedure and include dummy variable interaction with all the independent variables as dynamic regressors I find an error 'Singular matrix'. what should I do?
Hi guys, How do I incorporate the dummy variable in the ARDL?When I put my dummy variables in the list of fixed regressors, Eviews gives a warning "near singular matrix error". What should I do?
@@CrunchEconometrix Thank you. I just cannot hear what you did to the dummy variable to generate dum_sedu. Once I get that I will be good to go.Thank you for the video
@@CrunchEconometrix The only issue is when I interact the dummy variable with all regressors it is still saying near singular matrix but if I interact with variables less one it is bringing out results. Don't know why
hello, teacher i finished my data analysis but i got 2 problems the first one is i have 5 variables and dummy variable with 37 observations if i put the interaction term of dummy variable with another regressor i got in total 10 variables and the regression ARDL is not running so when i run the model with just dummy variable its ok the model is run . second problems is my cointEq coefficient is positif and significant and any test after this are ok .thank you
Yes, the outcome is expected because you'll lose degrees of freedom with a total of 10 variables and 37 observations. Solution: reduce the explanatory variables to 2, so that with the dummy and interaction terms you'll have a total of 5. Then re-estimate the model. So, simply use 2 essential explanatory variables simply because you are including dummy variables too. Keep me posted on the final results...
I really do not know how to interprete my results. Example: data from 1900 to 2010. I see breaks for first world war, second world war and oil crisis. I split my large annual dataset into smaller datasets => 1920 to 1980 with dummy 1 if [1940:1945] and 0 elsewere and interaction terms * this dummy Is this the correct procedure? How do I interprete this?
This is not a problem as there are tons of papers on dummy variables and which you must have reviewed some for your thesis. Take a look at their interpretations and adapt. You learn better and faster that way.
I have greatly learned from your presentations and want to thank you for all your efforts. I have one question: In my ARDL model, I have 1 dependent variable, 4 regressors; all variables used in the log form. Now i would like to add 1 dummy variable in the model, should I also transform it in logarithmic form or not.
Dear doctor I have another question...Normally, where we put a dummy variable, i use eviews 9??? in the box of dependent variable followed by list of dynamic regressors or in the list of fixed regressors? thanks for your availability
Don’t be disrespectful with your “Sir”. Why do you have to assume the person’s gender when it is obviously against what you said? It’s not only men who can teach. Also just don’t assume the gender!
Hi Nana, we offer consultancy on data analysis. Kindly email cruncheconometrix@gmail.com for the Terms of Engagement and consultation fee. Kind regards.
TH-cam 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, help me stay online.
Good evening prof, thank you as always.. I still haven’t found the video on how you determined and created the dummy variable 🙏🏿
You may want to watch other time series dummy variables variables. I showed the underlying data in the videos.
Thank you so much teacher for this videos . these videos really helped me . Thank you again for the attention that you pay on my comment and my question
Thank you very much. I wonder if you show us how to conduct Panel sub-sample or sub-region analysis in Eviews starting from spliting sample or creating region id. You may use Sub-Saharan Africa as case for this. Thank you in advance!
Yes, I could if EViews is code-driven like Stata.
Thank you Dr for this analysis. I have a question, the issue is I am using ARDL technique which is involving a dummy variable, everything is fine even the F statistic and the diagnostic tests BUT the problem is on cusum and cusum of squares test, it is not showing the two red lines and the blue line BUT rather it is showing a bar graph. IF i remove the dummy, the two red lines and a blue line on the cusum and cusum of squares will be shown but not when i use dummy. Also, is it ok if a person uses Ramsey test for stability when using ARDL instead of cusum and cusum of squares? Your response is appreciated.
Hi Takunda, the Ramsey test to the best of my knowledge is testing for omitted variables.
Prof, tq for the tutorial. Can you give the example of interpretation the coefficient of one dummy variable based on the result? Tq in advance Prof
Hi Wan, watch my videos on DUMMY VARIABLES, Error Component Models and Panel Sub-Samples. I explained the coefficient of dummy variables in those clips and I'm sure I explained in this video too. Thanks
@@CrunchEconometrix TQVM Prof
Thank you for this tutorial video. I would like to ask a question.
Why do you need to interact the variables with the time dummy? Do you have any book and or journal references on this? Thank you.
Hi Michael, I don't have any reference but you can always search online for textbooks or articles that modelled structural breaks. I only showed how it's done.
thanks prof, amazing video. but i have a question, when i try to testing unit root for variabel dummy (where, 0 before covid and 1 after covid), the result is "singular matrix". can you help me and what should i do? thanks before
Hi Jul, there's no need engage URT for a dummy variable. ARDL will automatically assign 0 lag to it.
@@CrunchEconometrix thanks Prof. so I don't have to do a unit root test, right ? thank'u so much prof
Hello, first of all I would like to thank you for your videos as they were very helpful for my studies and research. I would like to ask if there is a problem when one wants to incorporate a break that occurs in his last observations, meaning that there are not many observations that were affected. Can many "zeros" in the dummy variables (an example would be 95 out of 100 observations take the value zero) be a problem in a model (ECM, ARDL and so on)? Moreover, would it be logical if one tried to also examine how the lagged values of the structural break could affect a model (instead of the dummy variable itself) or there is not a sufficient interpretation of this?
Hi Leonidas, please reduce your query. Too long. Thanks.
Hello CrunchEconometrix, thanks for this amazing tutorial. My question is, if I run my bound test and the results shows no long run relationship between my variables, do i need to use dummy variables for my analysis if thier is still the presence of structural break? Lastly what are the adequacy test or diagnostic checks that I need to subject my Short run ARDL model to in order to know whether it is good for interpretation. Thank you.
Thanks Berk for the positive feedback...appreciated! Your queries are a bit confusing. Several unrelated approaches are muddled up. Please be more explicit. Using dummy variables may be specific to your study. I also covered diagnostics (hetero, serial corr, normality, stability) in my ARDL videos. Kindly watch them again.
Hello Sir. I have a question and if you can answer, I will be very happy for that. I have got two dummy variables in my ARDL model. When I put my two dummy variables in the box of dependent variable followed by list of dynamic regressors, it works. However, when I put my dummy variables in the list of fixed regressors, Eviews gives a warning "near singular matrix error". What should I do ? Thank you for your videos.
Hi Omer, it implies "multicollinearity". Including the dummies as a dynamic do not change your results because ARDL assigns 0 lag during estimation.
Hi professor, thank u so much for your video it is very helpful.
i am running an ARDL model with dummy variable for the break point specified by the chow and quandlt test for breaks. However I am wondering why the dummy coefficient is insignificant although the tests proved the presence of break?
Hi Ingy, it is possible for the break dummy coefficient to be statistically not significant. It implies that it has no impact on influencing the behaviour of the dependent variable.
While estimating var, why didn't you insert other variables under exogenous variables, min 2
Hi Ala, no need for inclusion of exogenous variables since lag is being determined for a particular variable.
Wouldn't the ARDL actual results be from the Long-Run Form and Bounds Test?
Alvin, yes if you are estimating the ECM.
thank you for this, i would like to ask you if we should proceed testing stationarity for dummy variable
Hi Achraf, there no point testing stationarity for dummy variable. Use them in their forms.
Thank you I appreciate that , I ask opinion from others and I research for my own and I found that dummy variable is stationary by nature in first difference, are this is correct ?
@@achrafbenssassi9946 I have since responded to this.
CrunchEconometrix so it is possible to integrate dummy variable in VECM after cointegration test
@@achrafbenssassi9946 You can check other online resources for that.
hi professor,
to assess the importance of each independent variable in determining the dependent should I use the standardized coefficients? or how to quantify each variable (or group of variables) contribution ?
Hi Ingy, I have no idea what you are referring to.
Brother can you explain the interpretation of the dummy variable in a little more detail please.
You may want to watch my specific videos on DUMMY VARIABLES.
Thank you for sharing these videos, i study a lot since follow your channel. May i ask you some questions? 1. can I use the dummy variables withe panel ARDL and panel VAR model?
Hi Xu, thanks for the compliments. Can't say exactly cos I've never tried it. But do further online findings on it, give it an attempt and see what you come up with?
@@CrunchEconometrix Thank you so much!
U're welcome, dear...please tell others too about my Channel!
Your link for the dataset used for this video is not working Dr.
Hi Kenechukwu, due to abuse I disabled the link. Some datasets are now available on my website upon payment of a token. Here is the link cruncheconometrix.com.ng/shop
How to OLS regression this ARDL model? While Eviews is making a parsimonious specification of it, I thought it would be better to regress like you did in the OLS approach video
Watch my OLS-based ARDL model for guide.
@@CrunchEconometrix with dummies I mean of course
when we use the long run and short run form, should the dummy variable be appeared on the short run and long run model?
Hi Adisu, this video clearly shows what to do. Kindly watch again and take note of the procedure. Thanks.
Hello, please i have a question can i use ardl model for sample of 180 observation or just available for smal simple. Thanks
Yes you can for time series.
Hi Dr Ngozi, I have a question related to control variable. My supervisor asked me to use inflation in my model as a control variable so how do I incorporate the control variable in Eviews. Thanks
Hi Mohammed, a control variable is an explanatory variable. Simply include it in the model as you did others.
@@CrunchEconometrix do you mean like independent variable ? Do you have any video to illustrate the use of control variable
@@MoonsUniverse140 Yes. Control variables are the same as independent variables. You give the same interpretation.
@@CrunchEconometrix thanks Dr Ngozi for your support, one final question on this, in Eviews help on control variable, they said to put !, So if I understood correctly, in the estimation equation, I put my dependent and independent variable and ! inflation because this is my control variable
Dear professor, thank you very much for another great video.
I have a question. When we write a paper, should we interpret coefficients of dummy variables as a short-run or long-run relationship?
Hi Djordje, interpretation will depend if at the end of the estimation the coefficient appears in the long-run and/or short-run parts of the estimation output.
@@CrunchEconometrix Thank you professor very much.
Guys, how did we get dum_lndc from dum? Then if we have say six independent variables, will we have dum interaction with the six independent variables? Thank you in advance.
Please watch the video again and jot some points. I made clear explanations while showing what to do.
@@CrunchEconometrix Thank you. I finally got the explanation. Lastly the dummy variable interaction can it be extended to 7 independent variables or they are too many?
@@CrunchEconometrix Must we do the dummy variable interaction for all independent variables or only the ones are of the main interest to us?
That will depend on your study objectives.
@@CrunchEconometrix My study looks at how mobile money has impacted on private investment. The independent variables are
GDP= the percentage change in real GDP;
GI = the ratio of public-sector investment to GDP;
CR = the change in credit to the private sector;
CPI = the percentage change in consumer price index;
TOT = the change in terms of trade;
RER = the index of real exchange rate;
OPEN = Openness of the economy expressed as exports plus imports over GDP
FI = Value of MPESA transactions
Elec_d = Election year dummy
e
When I follow your procedure and include dummy variable interaction with all the independent variables as dynamic regressors I find an error 'Singular matrix'.
what should I do?
Hi guys, How do I incorporate the dummy variable in the ARDL?When I put my dummy variables in the list of fixed regressors, Eviews gives a warning "near singular matrix error". What should I do?
Dennis, this video and others in its category shows you what to do.
@@CrunchEconometrix Thank you. I just cannot hear what you did to the dummy variable to generate dum_sedu. Once I get that I will be good to go.Thank you for the video
Interact the dummy variable with the regressors.
@@CrunchEconometrix The video was quite helpful. Have subscribed already.
@@CrunchEconometrix The only issue is when I interact the dummy variable with all regressors it is still saying near singular matrix but if I interact with variables less one it is bringing out results. Don't know why
hello, teacher i finished my data analysis but i got 2 problems the first one is i have 5 variables and dummy variable with 37 observations if i put the interaction term of dummy variable with another regressor i got in total 10 variables and the regression ARDL is not running so when i run the model with just dummy variable its ok the model is run . second problems is my cointEq coefficient is positif and significant and any test after this are ok .thank you
Yes, the outcome is expected because you'll lose degrees of freedom with a total of 10 variables and 37 observations. Solution: reduce the explanatory variables to 2, so that with the dummy and interaction terms you'll have a total of 5. Then re-estimate the model. So, simply use 2 essential explanatory variables simply because you are including dummy variables too. Keep me posted on the final results...
I really do not know how to interprete my results. Example: data from 1900 to 2010. I see breaks for first world war, second world war and oil crisis. I split my large annual dataset into smaller datasets => 1920 to 1980 with dummy 1 if [1940:1945] and 0 elsewere and interaction terms * this dummy
Is this the correct procedure? How do I interprete this?
This is not a problem as there are tons of papers on dummy variables and which you must have reviewed some for your thesis. Take a look at their interpretations and adapt. You learn better and faster that way.
Dear Ma'am, how u created dummy?
Hi Manu, I created it from excel since I'm using 0, 1 dummies.
What happens when the lag lengths are mixed for the variables say lag1 and lag 2?
Hi Afam, use the Toda-Yamamoto technique. Read more about the technique online.
@@CrunchEconometrix Prof, Is it Toda-Yamamoto technique are to analyze the direction of
causality? that test will apply before or after ARDL test?
I just told you to read up on the Toda-Yamamoto technique. Thanks
I have greatly learned from your presentations and want to thank you for all your efforts. I have one question:
In my ARDL model, I have 1 dependent variable, 4 regressors; all variables used in the log form. Now i would like to add 1 dummy variable in the model, should I also transform it in logarithmic form or not.
Hi Charifa, dummy variables are in 0, 1 measures. No need for conversion to logs.
Hi Charifa, dummy variables are in 0, 1 measures. No need for conversion to logs.
@@CrunchEconometrix Thank you Doctor
Dear doctor
I have another question...Normally, where we put a dummy variable, i use eviews 9??? in the box of dependent variable followed by list of dynamic regressors or in the list of fixed regressors?
thanks for your availability
Fixed regressors.
Sir, can dummy interactions have lag?
No...
Don’t be disrespectful with your “Sir”. Why do you have to assume the person’s gender when it is obviously against what you said? It’s not only men who can teach. Also just don’t assume the gender!
Dear prof pls make a video on NARDL
I've noted it Foday, I'll try...may I know from where (location) you are reaching me?
Can I get ur email? I want to send you a monthly Time series data so u help me with the analysis
Hi Nana, we offer consultancy on data analysis. Kindly email cruncheconometrix@gmail.com for the Terms of Engagement and consultation fee. Kind regards.