April 2016 Q5 - Performance Attribution

[GeorgeFS]

Hi,

I'm a little stuck with April 2016 Q5 which is a performance attribution question. There are a couple of places I'm struggling with:

1.
I am able to get the figure of 8.01% for the Fund return in 2012/13, but I can't seem to match the 11.50% for Fund Return in 2013/14 (I'm getting a figure of 11.60%).

To calculate Fund returns, for each asset I have calculated: (Start Price 2013 / Start Price 2012 - 1) * Holding %
... this gives the actual return for each asset, which I've then summed to give my actual return for the year
... as above this gives me the correct answer for 2012/13, but not for 2013/14 when I use the 2013/14 start prices.

To calculate Benchmark returns, for each asset I have calculated: (Start Price 2013 / Start Price 2012 - 1) * 10%
... multiplying by 10% as the benchmark assumes equal weighting in the 10 asset classes.
... then sum the benchmark returns for each asset class to give the total benchmark return over the year.
... when I do this I get 4.86% and 13.29% for 2012/13 and 2013/14 respectively.
... what am I doing wrong here?

2.
For parts (ii) and (iii) we are asked to calculate stock selection performance and sector selection performance.
... ordinarily, I would look to set up a Actual/Notional portfolio, consisting of actual weightings (ie the holding % column) and benchmark returns
... however, that would just equal the Fund return, since the benchmark returns are the same (use the same start prices).

I've tried using the formulae below, but that also doesn't get me to the correct answer:

Stock Selection = Actual Weight x (Actual Return - Benchmark Return)
Sector Selection = (Actual Weight - Benchmark Wegiht) x (Benmark return on asset - Overall benchmark return)

... feels like I'm missing something here with respect of how the benchmark is calculated.

Any help would be great, thanks!

 

[Colin McKee]

Hi,

I'm a little stuck with April 2016 Q5 which is a performance attribution question. There are a couple of places I'm struggling with:


This was a VERY difficult and long attribution. The ActEd ASET runs to nearly 7 A4 pages showing the calculation and mentioning alternative methods that may have saved time. I will try to give you a pointer, but it would be impossible to put full detail to this question in this forum (sorry).



1.
I am able to get the figure of 8.01% for the Fund return in 2012/13, but I can't seem to match the 11.50% for Fund Return in 2013/14 (I'm getting a figure of 11.60%).

To calculate Fund returns, for each asset I have calculated: (Start Price 2013 / Start Price 2012 - 1) * Holding %
... this gives the actual return for each asset, which I've then summed to give my actual return for the year
... as above this gives me the correct answer for 2012/13, but not for 2013/14 when I use the 2013/14 start prices.

I suspect you are getting your weights wrong for the actual fund at the start of the second year. If you do the first year correctly, you get (for equities A, B and C for example) returns of A(10%, 10%) B(-10%, 20%) C(20%, 20%). For the first year we get an overall equity return of (12.5/20)*10%+(0/20)*(-10%)+7.5/20)*(20%)=13.75%. But for the second year you need to work out how much we have in equities in total at the start of the year. The total fund at the end of the first year comes to 54.01, so in equity A, we have 15%*54.01 = 8.10, which after the second year performance comes to 8.1*1.1=8.91. Doing this for all assets makes the 54.01 at the start of the year rise to 60.22 at the end of the year. This is a return of 11.5% as per examiners report.


To calculate Benchmark returns, for each asset I have calculated: (Start Price 2013 / Start Price 2012 - 1) * 10%
... multiplying by 10% as the benchmark assumes equal weighting in the 10 asset classes.
... then sum the benchmark returns for each asset class to give the total benchmark return over the year.
... when I do this I get 4.86% and 13.29% for 2012/13 and 2013/14 respectively.
... what am I doing wrong here?

Not sure here. Your method seems OK. Our ASET takes 1/10 in each asset, weights by performance (in first year for equities overall we get 6.67% with 30% weighting, fixed income 6% with 30% weighting, commodities 4% with 30% weighting, cash 0% with a 10% weighting which gives average 5%. The benchmark rebalances, so the second year comes to 13.10% overall with 16.63% for the equity section for example.

2.
For parts (ii) and (iii) we are asked to calculate stock selection performance and sector selection performance.
... ordinarily, I would look to set up a Actual/Notional portfolio, consisting of actual weightings (ie the holding % column) and benchmark returns
... however, that would just equal the Fund return, since the benchmark returns are the same (use the same start prices).

Our ASET does something similar to what you say above. We create a fund with benchmark asset allocation, but use the actual returns. This is different from the benchmark/benchmark, and from the actual/actual. So you get an intermediate fund which you can use to strip out the components of return.

I've tried using the formulae below, but that also doesn't get me to the correct answer:

Stock Selection = Actual Weight x (Actual Return - Benchmark Return)
Sector Selection = (Actual Weight - Benchmark Wegiht) x (Benmark return on asset - Overall benchmark return)

... feels like I'm missing something here with respect of how the benchmark is calculated.

As I mentioned, this was a huge question and I cant give full detail here in a forum, but hopefully this sets you on the right trajectory.

Any help would be great, thanks!