Timeline for Residual autocorrelation versus lagged dependent variable
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 22, 2018 at 17:48 | comment | added | IrishStat | ................ | |
| Nov 22, 2018 at 17:27 | comment | added | Alexis | @IrishStat You have it precisely backwards: "The model that you are specifying can be restated as a particular case of a transfer function" The model you describe as a "transfer function" is a particular case of the GECM where $\beta_{x}$ has been constrained to be zero, meaning that your model assumes that the lagged effect of $x$ enters only in the equilibrium function. | |
| Jan 25, 2015 at 19:22 | comment | added | whuber♦ | @Irish If this answer is correct, then Alexis should not feel obliged to change the explanation or cast it into some particular form. You have frequently mentioned "transfer functions," and I think I have read all your (hundreds) of posts that refer to them, but I cannot recall reading any description of what they actually are. You might consider, then, posting an answer of your own in which you explain transfer functions and show how Alexis' model can be restated in those terms. | |
| Jan 25, 2015 at 18:07 | comment | added | IrishStat | why not ? If you constrain/specify a transfer function to a particular form you will the ECM . | |
| Jan 24, 2015 at 14:04 | comment | added | IrishStat | The model that you are specifying can be restated as a particular case of a transfer function just as an exponential smoothing model is a particular case of an ARIMA model. Please restate your model as a dynamic regression / transfer function. | |
| Aug 5, 2014 at 18:40 | history | edited | Alexis | CC BY-SA 3.0 |
Amplified the single equation generalized error correction model
|
| Aug 5, 2014 at 18:32 | history | answered | Alexis | CC BY-SA 3.0 |