CFAI Morning 2008 question 4 - Corner Portfolios

For anyone who has done this - they calculate the required return using he additive approach. If you use multiplicative, you get different weigfhts for the corner portfolios and the whole answer is wrong. Is there errate im missing? I dont see anything anywhere stating why they use additive and not multiplicative approach


If you use multiplicative you get a slightly different answer - and that does not even match the std. deviation expectations. That limit gets broken with the multiplicative way.

And there is no Erratum ever on any old exam.

So I guess it is a trial and error approach.

Maybe for Corner Portfolios go with the additive route.

yea guess so. Thanks for the response. Thats crap though they better not pull that in the real exam i dont understand why they cant just have a straight forward method I dont want to spend 30 min on exam, day having to redo a problem because i used multiplicative instead of additive approach!

can u show the problem/solution, not sure what ur talking about?


look at the 2008 problem and solution # 4 please.

paper is pretty much available.

Why would it be multiplicative? There’s no compounding. Return equals all position returns times their weight. Simple one period attribution. Only time I can think of same period returns being multiplicative is with currency, and that’s an special case bc there’s an interaction effect.



Think about it as of you have $100 $50 in cp1 w/ a 10% return = $55 $50 in cp2 w/ a 20% return = $60 Total Ret = $115 / $100 - 1 = 15% Same as .5(10%) + .5(20%) Only would compound if the returns were in sequential periods. Man, its hard to type this stuff on my phone…

the thing that complicates things in this particular problem is not any of the above factors - that it is an Pension Fund- and thus happens to be a LONG TERM INVESTOR (per Institutional IPS chapter). We also always do the Multiplicative approach, and they give us

With Additive - you end up with 0.25, 0.75 as the weights and an even 10% as Std Deviation (which is the max allowed).

With Multiplicative - you get 0.3, 0.7 as weights and 10.3% as the Std. Deviation. [Not sure if this would have been accepted, and there is nothing in the question to suggest that Additive was the way to go].

  • Thurlow Corporation is a U.S.-based manufacturer of skis and snowboards that began operations in 1995. In order to attract skilled labor, Thurlow offers employees attractive benefits which include a defined benefit pension plan and annual wage increases above the rate of inflation. An asset only (AO) approach to strategic asset allocation is currently used for the investment management of the pension plan. Tino Beveridge is a consultant to the board of trustees of Thurlow’s pension plan. The board asks Beveridge to recommend a strategic asset allocation for the pension plan given the following investment policy objectives:
  • Return requirement: Earn an average annual return of 8.7 percent plus management and administration fees of 0.7 percent.
  • Risk objective: A maximum standard deviation of portfolio returns of 10.0 percent.

When they linearly add the std devs, that is actually wrong math, but they note in the curriculum that it is just a way to estimate std dev. One of the many inconsistenties in l3

Cpk - your comments baffle me pften

Cpk - your comments baffle me often

Cpk - your comments baffle me often

Oh, I see what you are talking about. My bad. I’ve never done geometric or anything. Figures safer to go w/ lowrr

The way ive learned this stuff is to always use the multiplicative for institutional investors, this is the only time ive seen this happen its just worrisome. I wonder if there was an errata posted at some point, because the .75 vs .7 is a pretty signiifcant difference. I guess as long as we know the idea of how to solve it we should be good.

as I said before - CFAI exams guideline answers HAVE NO ERRATUM.

Even the other day someone sent a mail asking about the inflation before / after and received a rude email response back.

We are at their mercy out here :slight_smile:

I agree with the OP. When I first took it I used multiplicative. Doing so on the 08 ensures you break the max standard deviation. Good thing it was a mock and not the real thing. Lesson learned, because it changes your answer completely. It’s a bunch of crap. You’d think they’d leave some room there.

This is a clear fuck up IMO. I have emailed them on their multiple other fuckups, and got no response. I can only hope that they take this into account when grading, because typically when there is more than one answer the exam guildeline answers will show it, so this worries me because it makes no reference to multpile answers.

I think it is some theorem to do with Corner Portfolios .

They enter a realm of the weak force and all laws of physics turn upside down .

Stdev deviation : linear interpolation ,

returns : linear interpolations

… when it comes to corner portfolios only