i dont have the book in front of me …but dont we just need to explain the difficulties of multiperiod attribution (i.e.incorrectly accumulating attributes over time by simple compounding)?? or do we need calculation aswell?

Well that’s the $64,000 question with a lot of these calc questions.

I’m operating under the assumption that the exam is going to be a lot more qualitative than we are perhaps led to believe. But just to keep us honest, they are guaranteed to throw in a couple of calcs, even for an LOS that is not explicit about having to do so.

In any event, I figure if they throw one of the trickier formulas in our direction, most will probably get it wrong anyway, and they might take that into account in determining the MPS.

i actually prefer quantitative over qualitative (who doesnt :P) questions…i just hope they stay true to the LOS and not sneak in some calcs in ‘qualitative LOS’

Well there are calc questions, and then there are CALC questions. The former comprises bread and butter formulas that aren’t too hard to recall, even under exam conditions, while the latter is sadistic stuff like value a currency swap or determine the balance sheet CTA (from L2).

I find most of the formulas in L3 relatively intuitive with not much need for multiple steps, which made L2 such a bitch. The derivatives section is a notable exception, but I don’t plan to stress myself out unduly with that material. Would much rather drill down hard on stuff like private and institutional PM and get my points there.

I believe the text describes the pure sector allocation as the difference between the portfolio and benchmark weights, multiplied by the difference between that sector’s own benchmark and the portfolio’s benchmark.

this raises the question of what should be done when the question does not provide a benchmark specific to a sector, only one for the portfolio, or vice versa. in that case perhaps we would just multiply the difference in weights to whichever benchmark return is provided?

In case some of you have not got it yet, in fact both the formulas for (micro attribution) pure sector allocation and (global performance attribution) market allocation contribution are the same except for one small difference stated in bold below.

pure sector allocation: sum (portfolio weight for sector i - benchmark weight for sector i) X (benchmark return for sector i - benchmark return )

market allocaion contribution: sum (portfolio weight for sector i - benchmark weight for sector i) X (benchmark return for sector i)

In fact whether you put in the benchmark return figure or not does not impact your calculations. Both formulas will yield the same result, because benchmark return is a constant figure, sum (bechmark weight for sector i X benchmark return for sector i).

pure sector allocation = sum (portfolio weight for sector i - benchmark weight for sector i) X (benchmark return for sector i - benchmark return )

= sum (portfolio weight for sector i - benchmark weight for sector i) X (benchmark return for sector i) + [[sum (portfolio weight for sector i - benchmark weight for sector i) X ( benchmark return )]] (above term in italics sums to 0)

= sum (portfolio weight for sector i - benchmark weight for sector i) X (benchmark return for sector i)

can’t agree more with g3r41d but i think knowing what the formula stand for is better way, because some question could ask you to solve for only one sector pure allocation and you must use:

pure sector allocation i = (portfolio weight for sector i - benchmark weight for sector i) X (benchmark return for sector i - benchmark return ) if not , you will get the wrong answer :))

I realized today the I effectively skipped this section in my initial readings and viewings of the lectures and don’t really know it very well. I hope it doesn’t turn up on the exam.

Having reviewed it today, my conclusion is that memorizing the formulas is pretty useless as the exam will probably not exactly follow the formulas to a T.

Therefore its much better to understand what each of the 3 components plus the other 2 = 5 components are demonstrating. If you can do that then backing into the formulas as needed shouldnt be that hard, because we will prbably be presented with a simplified version in the form of a table or a few inputs.