Unit root tests in Eviews - Stationarity

[@attar8155]

Wow thanks, I got confused about KPSS but this video cleared things up!

[@JDEconomics]

Glad to hear it! Thanks for your feedback. Take care! JD

[@zinebkanche]

Thanks a lot for this video🙏🏻🙏🏻 I was so confused before, and everything is clear now.

[@JDEconomics]

Glad it helped! Thanks a lot! Feel free to subscribe and share my channel with your network. Good luck! JD

[@philipjavierocleda8889]

Hi, Just want to say thanks for the video, it was very helpful and your explanation is clear. 👍👍👍😄

[@JDEconomics]

Thanks Philip!

[@ALBERTMABUNU]

Thank you the video is very helpful

[@JDEconomics]

You are welcome!

[@alejandroruiz1525]

How can we determine when to include intercept, intercept & trend, or none in the equation of the ADF Test? Is it merely by looking at the graph or is there anything else?

[@JDEconomics]

Yes, by looking at the graph. However, you should always review the three specifications. Regards, JD

[@renatocanuto6746]

Hi. Thank you for the explanation! But I have a question! If LM-stat is smaller than the critical values, I don't reject the Null Hypothesis, so the serie is STATIONARY??? tY

[@JDEconomics]

Yes. You fail to reject the null hypothesis, therefore the series is Stationary.

[@renatocanuto6746]

@@JDEconomics Thank You.

[@EconomiaASAP]

Hey, can you explain how can I save this output as a variable to use in others regressions? I need to use Brazil's gdp for example, as a stationary variable in a regression, but I don't know how can I save the output to do it in Eviews.

[@EJOcio]

Hi, thank you for this. However, in ADF, what if the trend and intercept are both not significant, should I choose "None"?

[@JDEconomics]

Yes, you can select none. Regards, JD

[@Onlymagic00]

Hi, what can I do If results of the test are different from each other ?

[@JDEconomics]

Make sure that you are using the right specifications in the tests. Normally, you can visually see that if there is a trend, the series is not stationary. The tests and autocorrelation can give you some more info about seasonality, etc. Good luck,

[@fitrianisanusaputri5648]

Thanks before sir. But may i ask a question? Can we choose one of the stasionarity test to check that our data is stasioner? For ex: im just use the kpss to check my data is stasioner, without use another stasionarity tests like adf or dicket fuller. Or shall we have to use both of adf and kpss?

[@JDEconomics]

Hello! Thanks for your message. The KPSS test is used to test stationarity around a deterministic trend. This means that the KPSS test may not reject that the series is stationary, even if the series is increasing. In case you are wondering what is a deterministic trend, all it means is the slop of the trend doesn't change permanently through shocks. It can go up if there is a positive shock, but it will go back to the original trend then. Regards! JD

[@jurgencaceresbueno4023]

Hi, great video! I have a question I want to run a VEC model with 4 variables and one of them is stationary because I ran the ADF test and its Prob is lower than 0.5 in all the cases, and I was wondering if I can make it non-stationary and how. Also, idk if because 1 of the variables that I am planning to use in the VEC model is stationary I would have some problems or not, or if I can run any model at all. I would appreciate it if you can answer my question, because as far as I know all the variables of the econometric model must be non-stationary in order to run the model. Regards!

[@michelle86731]

Thank you so much for the very clear explanation! But I have a question. I do these three unit root test for a time series, and the result of ADF test and PP test is having unit root (unstationary, accept the null hypothesis) but the result of KPSS is stationary (accept the null hypothesis). I'm very confused about the result. Could you tell me how to explain this situation? Thank you!

[@JDEconomics]

Hi, Thanks for your message! I am happy to hear the video was helpful. Feel free to subscribe for me content! Explanation: If KPSS indicates stationarity and ADF indicates non-stationarity - The series is trend stationary. Trend needs to be removed to make series strict stationary. Once you applied differences, the detrended series has to be checked again for stationarity.

[@michelle86731]

@@JDEconomics Thanks for your quickly reply! According to the graph of the time series, I only include intercept in test equation at first and get the result that I mentioned. Are there other possible reasons for this result?

[@JDEconomics]

@@michelle86731 Hey, I'm sorry. If you can send me the screenshots of the things to [e-mail] I can take a look at it. Cheers.

[@michelle86731]

@@JDEconomics I already sent you email. Thank you!

[@emirarefa]

Hi thank you for your explanation And I have a question, if i am using trend and intercept in ADF test, is it influencing the method that I use? For example, here I am using ARDL, then, am I have to use ARDL model that include trend?

[@JDEconomics]

Hello Emira, Thank you for your message. If they have a trend, in the trend specification dropdown you can specify if the trend and intercept should be included. Kind Regards, JD.

[@emirarefa]

@@JDEconomics thank you for your answer. Do you have any literature recommendation for me to learn more about it?

[@zulekhaqadeer4147]

I have a question that while applying adf there is also t statistics and p values come but in kpss its just the p value... And if we want to write t statistics value in our work of kpss what we will do.

[@JDEconomics]

Hello Zulekha, thank you for your message. Unfortunately, the KPSS output only provides the p-values. The reported critical values for the LM test statistic are based upon the asymptotic results presented in KPSS table. In such case what you would do is base your analysis on the reported p-values. I hope that helps! If you liked the video, please feel free to subscribe for more content. Have a nice day! JD Econ.

[@fadzly8346]

hi, how to determine what's the level of significance based of adf t statistic and critical value?

[@JDEconomics]

Hi, I don't really understand well what you are asking. The critical values are based on sample size and significance level. The adf t statistics you use them to compare with the critical values. The critical values associated to a significance level, you have to find it using the statistics table. I don't know if that helps your question. Kind Regards, JD

[@musanuhuabdullahi3674]

how can it be stationary in the KPSS

[@JDEconomics]

Hello Musa, Thank you for your message. The Value reported in the KPSS test is not a P-Value, it's a statistic. The null Hypothesis of the KPSS test is stationarity, but your statistic 0.35 is in the rejection region (since it's bigger than 0.21; 0.14 and 0.11), so you will reject the null hypothesis of stationarity. Hope that helps, Best Regards, JD.

[@JDEconomics]

Hello Everyone! Thanks for watching! If you liked this video, please like and subscribe for more content! Your support helps me to create more video toturials. Please feel free to leave your comments and let me know if there is any topic you would like me to cover! NOTE: YOU CAN ACTIVATE THE SUBTITLES (CAPTIONS) AND TRANSLATE IT TO ANY LANGUAGE YOU WISH (IF DESIRED). Subscribe to my channel by clicking: kzbin.info/door/5P21WGFO4WRUlAiGLcwymg Thanks a lot! JD Economics.