“If we reduce the tracking error of the manager with the highest active risk, this is
very likely to reduce the plan-wide active risk of the overall portfolio.”
What if the benchmark were the broad market index ( i.e. investor benchmark=active manager benchmark ) ? Would your answer be the same?
If it was different , either you should qualify the answer or say the question is not correct because it could be different because the benchmark for style is not specified.
I guess this just got more complex for me. “if we reduce” is the big problem. Are we forcing it to reduce or is it happening by itself.
For all those who say that reducing active risk will lead to reduction in active return is not making sense. Active risk is the variablity of active return. What if the manager consistently earns the same alpha. As in there is a good active return but if the alpha is either positive or negative and constant, the active risk can be 0 as well though this is unlikely to happen. So If we reduce one manager’s active risk with no change in other manager’s active risks, isn’t the overall variability likely to be lower than before.
Yes, I think the “active risk” in the question is total active risk(for the manager and for the plan-wide), and the “tracking error” is the true active risk for the manager.
Correct me if I’m wrong.
So the question was about “true” active risk and not total active risk?
Is Tracking error = “True” acive risk = std dev of portfolio returns-normal benchmark?
If you execute a benchmark + 2% strategy with perfect precision your active return will be 2% but the variability of those returns are 0. If you reduce those active returns, your active risk would actually increase.
Fin , truong and serious are correct .
The ans. is True.
The q does not talk about active return , but the only way one can reduce tracking risk is by cutting active return . In any case that will reduce the overall portfolio tracking risk.
I can’t think of any scenario where the “highest tracking risk manager” even matters .
All that is important is that the weights ( allocations ) are unchanged and the active returns are un-correlated.
stingreye wrote:
Quick calc to prove:
active risk is a standard deviation so:
Mgr 1 Risk = 2%
Mgr 2 Risk = 1.5%
50/50 weight
so
[.5^2*.02^2+.5^2*.015^2]^1/2 = 1.25%
You can take out Manager 1 and risk actually increases becuase active risk is uncorrellated.
Your updated equation would be [.5^2*.00^2+.5^2*.015^2]^1/2 = [.5^2*.015^2]^1/2 = .75% so the active management did decrease.
you reduce active risk for one manager - you are also going to reduce the active return overall.
and as a result the active risk movement direction would not be known clearly.
I think it is FALSE.
发表于 2013-4-19 15:45
| 