I bet these guys used themhttps://twitter.com/tracyalloway/status/1364026615177519104 …
Good morning to everyone except people who use a sample covariance matrix in their portfolio optimiser.
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
ehhhh.... depends on your sample and depends on how much your optimizer relies on it.
-
Yeah ok, if you are not implicitly inverting it is probably fine. Or if n(observations) > 10 x n(assets) you might be alright too.
- Show replies
New conversation -
-
-
I'm writing my thesis on something related and I'd be curious about what is commonly used instead?
-
Lots of methods (eg shrinkage, clipping small eigenvalues) but most common approach I’ve seen is to use a covariance matrix derived from a factor model —https://twitter.com/macrocephalopod/status/1356731277337108482 …
- Show replies
New conversation -
-
-
Wait. So I'm not supposed to use the one that was printed in the instruction manual?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
It works. Especially weighted (e.g. EWMA) estimates. Depending on optimization and lookback can tune with shrinkage. Also not much benefit to factor models if you are already dealing with assets close to factors, e.g. diverse asset classes.
-
It can be ok if your portfolio is small enough. If you try to do this with a universe of ~3000 stocks you are in for a world of pain.
- Show replies
New conversation -
-
-
Honey! I always shrink my covariance matrices!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.