What's new

Question on 8-a-1 Grinold setup.xls

Thread starter #1
Hi David,

I was referring to the formula on slide 4 and 5 for the total risk and then to the formula used in the excel sheet 8-a-1 in C15 for the total risk.

The formula there is "=C2+C6+C11+C13", should it not be "=1+C2+C6+C11+C13"

Sudeep Manchanda

David Harper CFA FRM

David Harper CFA FRM (test)
Staff member
Hi Sudeep,

It recreates Grinold Table 4.1 (Ch 4 not assigned, just setup) so it is okay "as is" but i probably need to better label: they are monthly returns ... but thanks for noting the confusion b/c I can better label it - David
Thread starter #3
David I understand that it is a reconstruction from ch 4 (As a matter of fact this sheet helped a lot); but can you please explain the significance of the "1+" in the formula and how is related to return being monthly or annual.

I guess i would be able to understand what you are pointing to if I understand " why 1+ .. ?"


David Harper CFA FRM

David Harper CFA FRM (test)
Staff member
Hi Sudeep,

As you noted, my formula does *not* contain a “1+”

For example, I copied Grinold’s monthly riskfree rate of +0.26%
That would be equivalent to November * (1+0.0026) = December; or Nov * 1.0026 = Dec

(it’s beside the point, and more advanced, but notice that I can infer simple returns not log returns … log returns are time additive but not cross-sectionally additive .... but simple returns, vice versa )

…please advise if i got your point?

Thanks, David
Thread starter #5
I guess your point..... Just to be sure of my understanding let me take another example of a simple formula :

FV = PV (1+ r)^n ... the reason for the "1 +"is the same for the both the formulas.


David Harper CFA FRM

David Harper CFA FRM (test)
Staff member
Hi Sudeep,

I don't have a "1+" ... so i can't tell why you are asking for its rationale?
I am only showing the rate (r).

Because it's only a rate, it could be a simple annual (discrete) rate just as you show:
i.e., FV = PV (1+ r)^n is (simple) discrete annual rate

...but it could be used for a log return also
FV= PV*EXP(r*n), so that LN(FV/PV)*1/n = r as a continuous rate

...or for a discrete semiannual: FV = PV (1+r/2)^(n*2); until the compound frequency is clarified, could be any of these

thanks, David
Thread starter #7
Thanks David, The question is answered.

I would say that there are small connections (or links) between different reads which we have and with the help of "Bionic Turtle" these are made visible, this helps the overall understanding.