ARDL Bounds Test - 6of6 (Short Run Dynamics)

[@pawalucious89]

Very well explained, Prof. A question though out of curiosity. What is the theoretical justification for including the difference operator in the short run specification?

[@nguyentu3776]

thank you so much, your lecture helps me a lot

[@PatObi]

Your welcome Nguyen!

[@izzatabelle]

Dear prof, your video is valuable. at 8:06 when you test for short-run model of no-cointegration equation and using lag 1. Is it because of in the lag length test shows the optimal lag is in lag 1 ?

[@PatObi]

Yes

[@izzatabelle]

How if the result shows p-value above 5% whichs means no short run relation ? Can I solve this problem and make it shows significant relation ?

[@LifelongStudentBelgium]

Hi thank you sir for this really clear video. I want to ask, whether or not you have to include the current year as you did. Because you found optimal lag 1 for LARR. but in another video (maybe from a competitor: crunchmetrix) only the lagged differenced variables are used.

[@PatObi]

It all depends on the researcher.

[@ssshhhjjjj3134]

good question. I have the same doubt...it affects the stability condition

[@femitex]

Fantastic series bravo

[@arubawx]

Hi, just a question. Let say we are looking at GDP, Tourism arrival and crime. Were the idea is to get a picture on crime. If for example your ect is -0.85 and we know that there are causality running from long-run values of GDP,Tourism arrival towards crime, can we use the ect equation to see if either GDP or Tourism has a positive or negative relationship with crime? In the short-run dynamics its easy to see from the ARDL model, but the long run dynamics I am not sure. thanks marc

[@mdimranhossainmilon8223]

Dear sir, Thank you so much for your vedio. I would like to ask whether the result in table at 4:02, the value of speed adjustment ECT is positive. In that case what can i do to make it negative value.

[@Sarpamus]

fantastic videos. thank you

[@arritaberjani5264]

Dear sir, Thank you for your videos on ARDL, they were very helpful. I have a question regarding the short run coefficient interpretation. EViews selected the model ARDL(1,1,0,3,0). Does this mean that in the short sun there is relation between the dependent variable and the one with 0 lags)? This would help a lot. Thank you in advance.

[@PatObi]

It means that only the contemporaneous term of that regressor affects the dependent variable, if t is significant.

[@arritaberjani5264]

@@PatObi so, when I do the error correction form of the model those that have (0) are not included therefore, I have to look at the initial estimation of the equation and see if they are significant and then interpret the coefficient? Thank you so much!

[@Azam_Pakistan]

Great reised version. A few queries Sir, do you mean it is sufficient to explain only the contemporaneous term to report short-run causality and not its lags? Secondly if in a joint test these jointly don't turn out to be non-significant but the contemporaneous term is significant, can we interpret the contemporaneous. Grateful.

[@PatObi]

Thanks for your Q. In my view, all lags should be interpreted, including the current term if significant. Lags convey different meanings in terms of how long the effect of X on Y lasts. Keep in mind the current term (Xt) is part of the lag structure...it has a lag of 0. So yes, it should be interpreted, in my opinion.

[@Azam_Pakistan]

@@PatObi Thanks a lot.

[@smsm314]

@@PatObi Hello My Professor, The problem is that according to EViews 10 (and not EViews 9), this variable does not appear in the short-run equation !!!!!

[@owendeboer5525]

Hi Sir, Thank you very much for your informative videos. They have been of great help. However, I am currently still left with two questions. Namely, I have included an exogenous dummy variable in my ARDL model to account for a structural break which previously interfered with the stability of my model. However, when I try to conduct a new Cusum test, it only shows me the time period after the structural break. How can I resolve this? Additionally, what is the added value of assessing the optimal lag length of the ARDL model beforehand? Would it not be possible to select automatic selection, opt for max lag length (12) and let Eviews do the work? Curious to hear your thoughts. Many thanks.

[@chakraacharya409]

Obi Thanks ! I got ECT=-2.11 with statistically significance what happens! other serial correlation ,bound test, stability test are desirable.

[@PatObi]

Good. You can also check s.r. dynamics.

[@chelseajacob7272]

Hi, Thank you very much, that was a quick response from you. I really appreciate. However, from the video link you posted, it was on time series estimation, but my question is on estimating panel ARDL long run/bound test . I am using standalone Eviews 11 academic version and I could not get the long run/bound test from the dialogue box after clicking on the coefficient diagnostics from views to determine the F-stat. If the problem is from the type Eviews Ii am using, which model of Eviews do you recommend I should purchase. Thank you.

[@PatObi]

I use EViews 10.

[@21992840]

Hi, can I use the same procedure for panel data?

[@chelseajacob7272]

Hi Thanks was fantastic!!!!!!! Please I am having issues using Eviews 10 university Edition to estimate ARDL long-run /Bound test. I clicked on coefficient diagnostics in View, the dialogue box that came out does not contain long-run/bound test. What should I do please. Thankyou.

[@PatObi]

If that's the case, then consider getting the full version. Much easier. In the mean time, you can estimate the equation directly as I show in this older ARDL video: kzbin.info/www/bejne/aKWWmmyLe9-KodU

[@chelseajacob7272]

Hello sir, Thank you very much. God bless you.

[@chelseajacob7272]

Thank for your reply but I watched a video on it but the tutor used stata to explain it instead of Eviews.

[@amaimask8685]

Hello sir, thank you for your video it help me alot. I have done ecm and bound test and i have some question sir. In my bound test i find the data is cointegrated, however i didnt have negative and significant ect in ecm analysis, is that mean cointegrated result in bound test didnt always mean we have cointegration in the long run sir?, Is cointegrated result in bound test only mean we possibly have long relationship and ect result determine whether we have long run relationship or not?, If that so why some people always told the variable is cointegrated in the long run when the they find result in bound test or johansen test is good?

[@zoyashah7826]

Sir if my optimal lag selection criteria is 3 and I estimated short run coefficients taking lag 3.my ECM is negative and significant ..But all my variables current year value is coming insignificant but its lag values of previous years are showing significant.Then how do I interpret the short run dynamics??

[@angelaadomokhai4388]

Thank you for this. Please after running my ARDL ECM, I have only the constant and the error correction term, and no short run differenced variable. What does this mean please? Also can i still go ahead and explain the long run relationship since my f-statistics is greater than my I(1) bounds

[@PatObi]

If there are no issues with your data and model, your results may be indicating only a l.r. relationship. May be the variables have no s.r. dynamics.

[@pradeepdangi9614]

Wonderful.........Thank you

[@hanisamalin4366]

Hi, what is actually the meaning of the long and short run ? Is it the number of years or what? I still cant defined them properly, 🥲

[@chelseajacob7272]

Good day sir, Please when you have inconclusive result while estimating Bound Test, do you still estimate the short run ARDL model. Thank you.

[@PatObi]

Yes

[@chelseajacob7272]

Hello Prof, Kindly help me out here, I tried converting my variables into log form and I got an error message of "log of non positive number- Missing data generated" From what I read I discovered that this is as result of negative numbers in my variables. My question is - how can I convert negative numbers to log form in Eviews? Thank you.

[@PatObi]

Hi Chelsea, log of a negative number doesn't exist.

[@محمديوسف-ل6ع2خ]

very thanks,sir

[@smsm314]

Good morning Professor, Please Sir, why EViews 10, estimates the short-term equation without the variables having a zero lag? For example; ARDL(2,2,0) Eviews estimates the short-term equation without the third explanatory variable! I.e. we can not calculate the short-term coefficient for these variable? B.w.

[@PatObi]

The algorithm in EViews considers that to be the parsimonious model.

[@smsm314]

@@PatObi Despite we can calculate the short term coefficients of these variables!

[@smsm314]

@@PatObi A related question: is it correct to say that the variable with a 0 lag has no short run impact in the underlying ARDL model (due to the fact that it isn't included in the ECM regression, which is defined in the literature as the short run adjustment)? I.e. the interpretation of the short-term equation (ECM Regression) does not concern these variables (Zero lag variable), despite we can calculate the short term coefficients of these variables !

[@PatObi]

@@smsm314 You can do so, if you wish. But I believe EViews reparameterizes the model for parsimony.

[@PatObi]

@@smsm314 You can do so, if you wish. But I believe EViews reparameterizes the model for parsimony.

[@UshieSimeon]

Compliment prof, please give me the reference of this work

[@fawadbahij8871]

First and foremost, thank you very sir for the great videos and efforts you put. we do really appreciate your endeavors. plus, sir, if you don't mind can I please request you to send that PP slides through email to me if possible? if yes, I would really appreciate your cooperative attitude. I'm looking forward to hearing from you, very sincerely yours, FAWAD.