Anyway, in the 2012 AM exam, this would have given an answer inconsistent with CFAI’s answer, as they do the two step method (adjusting the value of the portfolio, then the beta of the portfolio), round each, and then add. Whereas S2000Magician’s method (more correctly) rounds at the end. In this example, you get an answer different by 1 contract.
Do I need to worry about this? You can even see in the guideline answer that CFAI have kind of screwed up because they’re acknowledging that you have to add together the amount of contracts you have to buy for each step, and if you use their unrounded numbers, you get the same answer that I got. It’s just if you round and then add, you get 1 contract fewer.
Can we see that as a general statement? In the answer keys to past exams they always round very early and mostly to one decimal point (if at all). Is there any topic where it might be a problem to round only at the end and leave percentages with 2 decimals? Of course, contracts need to be rounded at the end, cannot buy fractions. But things like required returns, allocations or payments?
Works for all durations and betas as long as you correctly specify the dollar duration and dollar beta of the target, the current portfolio, and the forward contract.
So if the cash position has a 0.25 duration, we must take that into account in whichever part of the formula (portfolio, target, or forward) the cash position is relevant for.
I have a question related to this question. Why we need to adjust the beta of the portfolio. Should we use the notional principal of the original portfolio or including the synthetic one too ?