micro attribution question

[Gmax]

Should the sum of these 3 returns = difference between portfolio return (sector) - Benchmark sector return
Pure Sector: (Wp - Wb)*(Rbj - Rb)
Allocation Selection: (Wp - Wb)*(Rp - Rb)
Within Sector: Wb*(Rp - Rb)

 

[ajb1]

Yes

 

[Gmax]

ajb1 wrote:
Yes

It doesnt work on this question… Can someone explain why?
from 2015 AM Q-5D
Financial sector data: Port weight 31.35% Becnch. weight 11.79% Port fin sector return 6.82% Bench fin sector return 4.98 % Total Bench. return 4.01.
Pure sector allocation = (.3135-.1179)*(4.98-4.01) = .1897
Within sector selection = .1179 * (6.82-4.98) = .2169
sector selection/allocation = (.3135-.1179)*(6.82-4.98) = .3599 vs. expected 1.84% (Port sector return 6.82% - Bench sector return 4.98 %)
Total : .1897 +.2169 +.3599 = .7665 %

 

[cpk123]

i do not think your surmise is correct. do the algebraic expansion of each of the individual items - and see if you arrive at what you wanted.
One factor is rbj the other is rb - which are both different.

 

[sharky7]

Also check for the rounding, sometimes is a source of discrepancy in these questions.

 

[Gmax]

tried everything, it doesnt work for 2015 AM Q-5D . Please point out there is my error ?

 

[ajb1]

Apologies, I didn’t understand your question. The sum across all the sectors (Financial, Energy, etc.) when broken down into these three parts should be the portfolio return - benchmark return.
The reason you can’t sum these and get to the 1.84% is because the 1.84% is the excess return within the sector, but this micro attribution analysis is including the impact of different weightings.
To answer your question again, no, the sum of these parts do not equal the excess performance within the financial sector.

 

[Gmax]

ajb1 wrote:
Apologies, I didn’t understand your question. The sum across all the sectors (Financial, Energy, etc.) when broken down into these three parts should be the portfolio return - benchmark return.
The reason you can’t sum these and get to the 1.84% is because the 1.84% is the excess return within the sector, but this micro attribution analysis is including the impact of different weightings.
To answer your question again, no, the sum of these parts do not equal the excess performance within the financial sector.

Thank you! i start doubting my simple algebra skills