Timeline for Producing hedge ratios via regression via returns and not price
Current License: CC BY-SA 4.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 21 at 13:07 | comment | added | KaiSqDist | Hi, if you feel my response has helped you, you can give an upvote (if you haven't done so) or accept it as the solution (if it answered your question). :) | |
| Apr 14 at 8:37 | comment | added | KaiSqDist | If the residuals are homoscedastic, then it would be okay. If they are heteroskedastic, there are methods meant for that to treat the errors, such as the nonlinear OLS/Kalman filtration methodology for option pricing model calibration. On the price regression, I am honestly not sure, because I haven't worked much with them. | |
| Apr 13 at 22:29 | comment | added | ChairmanMeow | Thank you for your answer, and it intuitively makes sense to measure two assets (however different they are) via change vs. change, or returns. Do we simply just discard the constant from the model as a way of saying that's errors or unexplained part and just use the beta coefficient? Also, as mentioned above sort of a related question. If we were to take asset x and use regression to imply a fair value for asset y, then would using price be ok? | |
| Apr 13 at 21:42 | history | answered | KaiSqDist | CC BY-SA 4.0 |