Cointegration tests

[@Name-dw5vk]

Thanks you so much! Really helped me a lot for my econometrics by clearing my doubts as my lecture notes from my undergraduate course are not as detailed. Just wanted to express my gratitude to you as I’m sure it must’ve took a lot of time to be recording these videos. Once again, thank you!

[@janetjock9862]

Thank you for all of these videos! They have REALLY helped me understand all concepts in stats!

[@nanastuart8436]

Thanks a lot Ben. You are not only smart to learn for youself, but also smart to teach for the others.

[@yourdeadmother]

ben you have many videos, perhaps you could consider arranging it in a ordered playlist so we can watch it one after the other in a meaningful way!

[@SpartacanUsuals]

Hi, thanks for your message. If you go to my channel homepage all the videos are arranged into playlists. Hope this helps! Best, Ben

[@Andrei118]

Bless you for this wonderfully simple explanation

[@veronikaberezhnaia248]

I join to other thanks! I have a job project to be done urgently, and your videos help me to grab the concept

[@miamu7593]

Thank you, Ben. This is really helpful. Your explanation is clear and logical. Thank you.

[@SamuelTanzw]

OMG thanks so much i'm like a noob in stats and have been trying to decipher the 2nd regression needed for co-integration

[@MyMpc1]

Man, you're brilliant.

[@DimitriBianco]

For testing t < DF2 < DF1, would you use the augmented dickey-fuller?

[@charlie3k]

didn't expect to find you in this comment section :P

[@Helena-rq1qx]

This was a very helpful explanation, thank you!!

[@dexenxi812]

Hi Ben. Your video helps me alot!. In your example, yt and xt are both non-stationary. Can two stationary time series be cointegrated? If no, how can we test whether the two series are varying "similarly"? Thank you.

[@amaimask8685]

Thank its very nice video ben, so it that mean if there is cointegration relationship between variables we can use our ols regression output?

[@vivekbanerjee9653]

PLEASE ADD VIDEOS OF ARCH,GARCH MODELLING AND ESPECIALLY OF BIVARIATEGARCH MODELLING. THANK YOU

[@robbaggins]

Hi Ben These videos are brilliant - they're getting me through my dissertation right now. I have one question on this particular video though, is the test that you are describing the same as the Johansen test, or is this a different test for cointegration? Thanks Rob

[@znanyun]

This seems Engle & Granger test.

[@wen4205]

Brilliant explanation, thank you so much Ben!

[@emiliano8302]

Why are you including a constant when you test residuals with the ADF test?

[@bobo0612]

Hi! I have a question at 3:33, if we reject the null, then delta does not equal zero, which means as delta(u_{t}) changes as u_{t-1} changes, how come it is stationary? Shouldn't LS:delta(u_{t}) be a constant in order to be stationary? Thank you in advance for my question!

[@jobiraj2592]

Hi, If Johansen cointegration test confirms that an explanatory variable has a significant long term impact on the dependent variable, should the variable also have a short term significant impact when estimating the vector error correction model? If normalized cointegrating coefficients are insignificant, can we still conclude that there is a long term relationship because we said there is cointegration first? What are adjustment coefficients? are these the ones we report as our long term coefficients? Thank you!

[@Azam_Pakistan]

A lot of debate going on at the Hossain Academy Facebook page that the dependent variable can not be I(0) in an ARDL model for bounds testing approach as cointegration can not be there. Your views please

[@santiago-9074]

what about the cointegration if i run the unit root test for the first time on "Xt" and "Yt" and the two both are non stationary besides that "Vt hat" is I(0)?

[@mrcholo4604]

Thank you Ben!!

[@JMRG2992]

I've got a question. Lets say i have X_1 and X_2 variables. Running tests it was shown that X_1 is stationary and X_2 is not stationary. So X_2 has been transformed to be stationary using growth rates. Now Y has only stationary variables to be defined. Y= X_1 + X_2. Doing the Granger test of causlity is meaning ful here? ot it is only plausible if X_2 without transformation has a unit root. and seem to be related to Y. Or i can either test Granger, and cointegration with variables that are non-stationary?

[@KrishnenduJ-hc5fg]

This is the Engel & Granger test, right?

[@31896eneri]

I it possible for OLS and cointegration to hae different signs?

[@DistortedV12]

Isn't this like a conditional independence test?

[@domingosnhamussua3070]

and what about the johansen test?

[@nicolaswang7181]

Thank you for uploading so much helpful videos about econometrics, and by the way how come the topic that our lecturer need a whole week to explain and u just take one or two five-mins video???

[@SelvesteBjorn]

Hi Ben, great videos - thanx! At 5:10, what are you writing to show that the embedded Dickey Fuller test is "that much more negative" than the standard?

[@PaulJohn204]

-ve = negative

[@yiwanglim6624]

Hi Ben, i've got a question and was wondering if i could email you about it. Many thanks!

[@smsm314]

Hello My Professor, Please Sir, If we have 5 variables I(1), 1 variable I(0) and 1 variable I(2), how to do the cointegration test? Cordially.

[@Vale01-b1x]

An example with real numbers would help a lot

[@languangeme]

Hi Ben Thanks so much for your videos. I have been using them to help me study for my econometrics class. I am wondering what would you do after testing that two series are indeed cointegratable. How do you report the result and the long-run relationship between them? Is it just the e term?

[@aidanhorn]

I've only just started learning about this, but I assume you report the p-value of δ1?

[@tanushreedutta960]

Why do we need to delve into cointegration. Can't we simply transform the non-stationary time series into stationary time series . When it's I(1) it doesnt seem to be a problem. So why are we doing this?

[@znanyun]

When turning I(1) to I(0), lots of information will be lost in transformation. The linear relationship might only exist between those two I(1) series.

[@evanjones8373]

Hi Ben, you mention a stricter version of the Dickey-Fuller test statistic, where is this published?

[@richardhavell4492]

Hi Evan, I know this is two years old but the stricter Dickey-Fuller test stats are on page 766 of the Hamilton (1994) textbook and in the Appendix of Phillips and Ouliaris (1990) in Econometrica, Table II: finpko.faculty.ku.edu/myssi/FIN938/Phillips%20%26%20Ouliaris_Asymp%20Props%20of%20Resid%20Based%20Tests%20for%20Coint_Econometrica_1990.pdf

[@okechukwuobiora8155]

Hi Ben, this is very helpful. Thank you. But please, can you do a video on multivariate cointegration?

[@abdullahilmamun6301]

Very nice, so useful, thanks a lot

[@LutzSoedibjo]

BRAVO

[@SpartacanUsuals]

Thanks!

[@piperni9510]

I love this video!

[@santiago-9074]

What if Xt~I(0) Yt~I(1) Vt_hat~I(0) ?

[@clusterb8gaxilulu]

Thank you Ben! For all your videos that save my ass :D Thank you!

[@lastua8562]

Why the last sentence?

[@lastua8562]

"amend" means prepare a new version of, i.e. use the different DF

[@hojunseo1210]

Too many 'it turns out'..s