# Barra covariance matrix construction

I am trying to replicate the covariance matrix used by Barra risk models.

All Barra models have half life parameters for volatilities and correlations (e.g. if the half life for volatlity is 90 days, the decay rate is $$(1/2)^{90} \sim 0.991782$$).

And there is Newey-West volatility and correlation lags with $$n$$ number of days. Does anybody know how do these Newey-West adjustments work within Barra framework? Another parameter that I am not sure of is half life for volatility regime adjustment. How is this incorporated in building covariance matrix?

Thank you!

• I used to wonder about the details also and barra used to send us ( compan I worked for was a client of barra ) info on it but it was never close to the level where it could be re-produced. I don't know if they still do but, at the time, barra was selling their risk models to many, many, many companies and making quite a fortune so I'm confident that the information you seek is not available publicly and Barra will most likely not provide it. You couild ask them what they do provide and obtain that atleast. – mark leeds Nov 3 '18 at 5:19
• Here's something at the link below but it won't be enough. Otherwise, whatever company they are now a subsidiary of ( I guess alacra ) would not want them as a subsidiary. alacra.com/alacra/help/barra_handbook_US.pdf – mark leeds Nov 3 '18 at 5:22
• Yeah Mark, they just provide with cov matrixes for their different risk models with next to zero transparency. I was just trying to see and reproduce the cov matrix since they do give the historical returns for the factors. That paper that you linked and the one that is still around publicly are quite outdated. The Newey-West adjustment, I believe, kicked in a couple years ago or so ... – AK88 Nov 3 '18 at 22:15
• I mean there should be some people/companies out there that try to validate what Barra is producing for risk modelling. While I understand that it is difficult, I hope not everybody is taking the numbers as they are (without checking/verifying that is). Maybe I am wrong ... :) – AK88 Nov 3 '18 at 22:17
• Oh okay. I haven't dealt with barra since the mid-90's. I used to go to the conferences so I remember getting excited about the hopeful transparency and then being pretty dissapointed. Most not but there may be a few that try to back it out, but, if they were successful, they're probably not going to say much about the details either. Richard Grinold and the other author of that active portfolio management text ( whose name escapes me ) are the people you should try to connect with because they were there during the earlier years of Barra, although no longer I don't think. Good luck. – mark leeds Nov 4 '18 at 23:50