Timeline for VAR model for first differences (not a good idea?)
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 18, 2018 at 18:03 | history | edited | Jacob H | CC BY-SA 4.0 |
deleted 6 characters in body
|
| Sep 17, 2018 at 21:38 | comment | added | Jacob H | Updated, see above | |
| Sep 17, 2018 at 21:37 | history | edited | Jacob H | CC BY-SA 4.0 |
added 1610 characters in body
|
| Sep 17, 2018 at 15:39 | comment | added | NPHA | How does that change the situation? | |
| Sep 17, 2018 at 15:39 | comment | added | NPHA | @JacobH Lets say it’s forecasting. | |
| Sep 16, 2018 at 4:53 | comment | added | hejseb | @Kplusn: You just need take a look at the VECM (eg here: en.m.wikipedia.org/wiki/Johansen_test) and note that a model in first differences effectively ignores the long-run part ($\Pi X_t$). Obviously, by ignoring this part you are losing something. | |
| Sep 16, 2018 at 4:42 | history | edited | Jacob H | CC BY-SA 4.0 |
added 1649 characters in body
|
| Sep 16, 2018 at 3:49 | comment | added | Jacob H | Ok, well what is your goal? Forecasting or Inference? | |
| Sep 15, 2018 at 7:09 | comment | added | NPHA | Your answer pretty much goes trough things that I already know. Im looking fir explicit example of the situation when the information is lost. In other words I would like to see two time series $x_t=...$ and $y_t=...$ where I could see that taking the differences something is lost explicitely. The math is more powerful for me than just the words. | |
| Sep 14, 2018 at 23:01 | history | answered | Jacob H | CC BY-SA 4.0 |