Timeline for Why do we need a VECM specification if the I(1) processes are cointegrated?
Current License: CC BY-SA 4.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 15, 2019 at 7:33 | comment | added | Young | Ah never mind. Omitting such error correction term would make the lagged differences be correlated with the error term. I got it, thanks. | |
| Mar 15, 2019 at 7:29 | comment | added | Young | Thanks a lot. Would this mean that the expectation of the error term vector in the VAR in first differences is non-zero, though covariance-stationary? I understand that the estimates will be biased in that case, but how can we show that the estimates are inconsistent as well? | |
| Mar 15, 2019 at 7:27 | vote | accept | Young | ||
| Mar 25, 2019 at 17:31 | |||||
| Mar 15, 2019 at 5:11 | history | answered | dlnB | CC BY-SA 4.0 |