In the 2008 morning session question 4 , part B:
Why would we use leverage only to the point of hitting the return requirement? The answer solves by finding the portfolio allocation equal to the return requirement but that leaves us under the max standard deviation. I answered the question by maxing out standard deviation to 10%, which results in an expected return close to 9.55% – further this is greater than the answer’s return of 9.399%
anyone else agree?
did your answer create the portfolio with the highest sharpe? That is the key.
I do not think the book every suggests maximising the std dev is the right approach. you need to meet the return requirement … with which you get 0.25 CP3 and 0.75 CP4.
and for this combination -> std dev = 10 (it is the weighted avg of the std devs of the portfolios).
how did you get a different answer of 9.55% as you have stated above?
I had another related question: (Calculation of Sharpe ratio for the Combination)
their calculation of sharpe ratio is 0.25 * .46 + 0.75 * .51 = .4975
however we know the combination portfolio has a return of 9.4, std dev = 10 rf = 4.5
so (rp -rf)/stdev = 0.49 < 0.4975 calculated by them.
which is right?
There is no end to the variations.
you could take the tack that StdDev of a two-asset portfolio where the assets are completely uncorrelated should be Sqrt( ( WtA*stdevA)**2 +( WtB*stdevB)**2) .
But does the asset allocation chapter do this ? noooooo ! They do it as WtA*StdevA+WtB*StdevB.
If you can do that with risk , you can pretty much do it with anything else , right ? So Sharpe = WtA*SharpeA+WtB*SharpeB is the right answer if you get an asset allocation question!
Never mind , I got that clarified in the text . The Sharpe can be approximately calculated as the linear combination of the sharpe’s of the two corner portfolios. The text says that the stdev should be calcuated using weighted variance and covariances but can be approximated by a linear combination of the stdevs of the two adjoining portfolios.
Any how I have not yet checked the guideline answers
I have a follow-up question to this:
ii) So the right approach is to first calculate the weights that satisfy the return objective, 25% and 75%, and then after check that this does not give a standard deviation above 10%?
But what if, lets say this would result in a standard deviation above 10%. What would then be the correct way to solve the weights, should we then first have solved for the standard deviation?