Books with examples on Attribribution analysis/Performance Measurement.

[Canuck_Act]

Hello.

I was wondering if anyone knows of any good books, which could be used to practice calculations for performance management, especially attribution analysis. I have no problem when the attribution analysis only concerns two dimensions (i.e. Stock selection and asset allocation) and even 3 dimensions, if we include currency.

However, I cannot find books or even algorithms for more general scenarios or with such examples, so that I can practice.

Furthermore, I have noted that some CFA methods in the simple two-dimensional case, arbitrarily attribute to an “interaction factor”, in addition to the stock selection and asset allocation attributions, so I am not keen on adopting their methods as I don’t like making arbitrarily attributions to factors which the investor cannot control. Since the attribution is arbitrary, why not allocate it to one or the other (or both)of the stock election/asset allocations values?

Thanks in advance.

Cheers.

 

[Colin McKee]

I don't actually know of a book that I can say uses the SP5 methodology, nor am I certain that the SP5 examiners use a particular methodology from a particular text when they write questions. I suspect that they take past paper examples and tweak them a little. So I am struggling to help on this one. You say that you are generally happy with attribution to stock and sector selection, and even happy when the currency has to be split out as well (I think this has only happened on one occasion). So my question is whether you can provide an example of the sorts of scenarios (maybe from past papers?) that you have found difficult in the past?
In general I agree that these can be hard, particularly in an time-constrained, exam situation. But there are a great many examples in the past papers going back to as early as 2005. I have found that most exam questions resemble something that has popped up in a past paper before, so those would always be my first recommendation.

 

[Canuck_Act]

Thanks for your reply.

So, in practice, do most large institutional investors only consider attribution analysis on two dimensions (asset/sector allocation and stock selection) or 3 if we include currency? Perhaps that is the reason why the IFoA and other educational bodies do NOT cover the matter more thoroughly?

My question was also aimed at increasing my personal knowledge, as there is an open job vacancy at the only local potential employer of people with actuarial/investment knowledge and skills, but since it is a very large institutional investor, I worried that I might not have sufficient knowledge to carry out the analysis properly.

 

[Colin McKee]

It is hard to answer for large institutional investors, because I suspect they all have their own methods of performance reporting and attribution, which will all be very different. The important thing about the IFoA material is not that it replicates what happens in particular institutions, but that it gives you the tools to (given the time) analyse and understand any system used by any institution. My own experience in a fund management organisation was that much of the official performance work (both relative to other institutions and relative to benchmarks) was outsourced to a specialist provider of such information, who took our raw data and provided regular performance reports, placing us in a league table of similar funds and giving us very full attribution. The methods used were proprietary, but I suspect would have been disclosed somewhere. We also had a less sophisticated in-house system which gave us approximate performance information on a daily basis, but relied on daily valuations which were a bit more subjective in places. It would have taken many weeks to become an expert on either of these systems.