I saw the following formula to calculate the covariance between two indexes: Cov i,j= Bi*Bj*Variance of the market where: Bi= Beta between the market and portfolio I Bj= Beta between the market and portfolio J Cov I,j= covariance between portfolio I and J Does this formula look familiar to anyone? Where in the CFAI notes is it? Thanks.
looks familiar but I certainly can’t find where it is in the CFAI… Schweser man at heart
it was in level II last year (CFAI books)
Someone has posted this very formula a few days back. This is probably referred in one of the sample questions in CFAI.
I believe this formul was in L3 2007 curriculum in SS3 “Quantative Methods”. I can’t remember I saw it anywhere in the curriculum this year.
ahb Wrote: ------------------------------------------------------- > I believe this formul was in L3 2007 curriculum in > SS3 “Quantative Methods”. I can’t remember I saw > it anywhere in the curriculum this year. I was wrong . Check this LOS 23.c / Book 2, p. 70 (SS7 Economics) in Schweser notes
It is in Capital Market Expectations reading (I think 23 but not sure - somewhere around page 65) in Volume 3. It comes from the discussion of multifactor models under statistical methods used by the analysts to develop forecasts. The formula you list above is the case when there is only one factor driving the asset class returns.
“The formula you list above is the case when there is only one factor driving the asset class returns.” That’s probably true with enough assumptions added. The formula above is not true in general. Schweser seems to have an institutional misunderstanding that if r(1,2) = x and r(2,3) = y that r(1,3) = x*y. I’ve corrected that for them a bunch of times, sent them examples, asked them to try a proof, etc. and they just keep printing it and telling people it’s true. I think I would worry about it more except that in some sense a whole generation of misinformed financial analysts is not necessarily bad for me.
This is a formula in SS7 followed by the formula for calculating risk premiums for asset classes in a global context.
The formula is in the volume 3 of CFAIII book. And, the formula is the one part of the market model mentioned in CFAIII book.
Well, it’s probably true with “The formula you list above is the case when there is only one factor driving the asset class returns.” as I said above, and whatever else they mean specifically by that.
I was never big on deriving these things, but it looks plausible if you are assuming that whatever isn’t correllated with the market factor is purely random and uncorrelated. It wouldn’t work, for example, if asset A and B were stocks in the same industry, since deviations from the market*beta return would probably be correllated as well.
bchadwick Wrote: ------------------------------------------------------- > I was never big on deriving these things, but it > looks plausible if you are assuming that whatever > isn’t correllated with the market factor is purely > random and uncorrelated. It wouldn’t work, for > example, if asset A and B were stocks in the same > industry, since deviations from the market*beta > return would probably be correllated as well. That’s probably about right. I’ll check it out when I feel like it.