I’m a bit confused about the attribution analysis component “security selection”. I’m assuming in global performance attribution, this is “security allocation” and in micro performance evaluation, this is “within-sector selection”. First of all, is this correct? If so, why is the global formula weighted by portfolio weight, and the micro formula weighted by benchmark weight? Any thoughts on this would be appreciated. I’m trying to understand the formulas so memorization isn’t so bad. Thanks!
In “within sector selection” you are implicitly assuming that benchmark weights are equal to the portfolio weights. So that you are able to attribute the returns to security selection.
picnic Wrote: ------------------------------------------------------- > I’m a bit confused about the attribution analysis > component “security selection”. I’m assuming in > global performance attribution, this is “security > allocation” and in micro performance evaluation, > this is “within-sector selection”. First of all, > is this correct? > > If so, why is the global formula weighted by > portfolio weight, and the micro formula weighted > by benchmark weight? > > Any thoughts on this would be appreciated. I’m > trying to understand the formulas so memorization > isn’t so bad. > > Thanks! Yeah I just noticed that the two have differing weights. I don’t know why it is just that it’s another fact I need to memorise!
picnic Wrote: ------------------------------------------------------- > I’m a bit confused about the attribution analysis > component “security selection”. I’m assuming in > global performance attribution, this is “security > allocation” and in micro performance evaluation, > this is “within-sector selection”. First of all, > is this correct? > > If so, why is the global formula weighted by > portfolio weight, and the micro formula weighted > by benchmark weight? > > Any thoughts on this would be appreciated. I’m > trying to understand the formulas so memorization > isn’t so bad. > > Thanks! It is perfectly explained in some other post by someone (drkstar?) but I am having difficulty finding that post. The poster summarized the logic behind the formula very nicely. I’ll keep looking and let you know if I find it; which will help you understand why it is portfolio weight in global attribution but benchmark weight in micro attribution.
That is perfect! Thank you SO much! The reason is because the global formula doesn’t have a separate selection/allocation interaction term, so it is combined in with the selection term, so you use the Wp not Wb.