News
eFocus - Autumn 2007
Performance Measurement – More Compromise than Rocket Science
Industry Comment by Paul Copplestone, Gartmore.
This article has been written with reference to Spring’s Performa eFocus newsletter article written by Peter Ellis of Investit. Whilst we would agree with many of the points he makes, there is an opportunity to “add a bit of colour” to a couple of the themes that Peter draws on.
Peter states that, “The real challenge facing performance teams is not the theory of performance measurement but the practice”.
The theory part can be covered later but first the practice.
Practically every article that has ever been written about this area contains the words Data Integrity, yet little of what has been written ever offers any advice as to how to articulate this problem to senior management in a way to get their attention. Try this;
“The more accounting noise there is in your returns, the higher your volatility and the lower your Sharpe and Information ratios will be.”
We have found that this phrase usually gets a CIO’s attention and it has even been known to motivate them to do something about it!
As Peter states, a lot of the data issues we face surround the need for daily stock and fixed income attribution, but there are ways around this.
About ten years ago, I sat in a meeting where a number of Heads of Performance were earnestly discussing the need for daily stock level attribution. One very articulate and experienced individual was of the view that analysis of how many basis points one stock had on yesterday’s return would not (and indeed should not) influence a fund manager’s decision.
At the time I agreed with the view, but then a couple of years later the company where I work started running long/short equity hedge funds and aggressive long only funds.
Given that the fund managers now very much want to know the impact of each stock, not only on a daily basis but in real time, as the holding period of the positions can on occasions be less than one day, we quickly learned to compromise. We implemented an absolute attribution system for the hedge funds that includes all transaction costs, but moved to daily buy and hold attribution for the long only funds. This analysis is available to them as of close of business the previous day, and works in conjunction with their own desk real time systems that tell them how they are doing today.
We have had any number of discussions with performance professionals as to why they do not feel comfortable with the buy and hold approach. The answer is simple;
Data Integrity can be very expensive. Sometimes you have to be prepared to compromise in order to supply information that is cost effective. Concentrate your resources on getting the fund returns right and don’t get too lost in the need to attribute down to basis points.
Too many performance system implementations end up overdue and over budget; this is damaging, not just to your career but to all of us.
Which leads nicely onto Peter’s comment on the theory of Performance Measurement;
I’d guess that anybody in the industry who knows me would be aware that I don’t much go in for theoretical debates, and do agree with Peter’s sentiments, but one should be aware that in the right context these debates can be very valuable.
It is important to have a good understanding of the issues that these debates are about, which in turn makes you more aware of the implications and potential pitfalls of whatever solution you choose to implement.
The information you can glean from these debates can become invaluable if you face the need to make compromises in order to make your solution fit the budget that has been made available to you.
For business and professional investors only. This article should not be circulated to private investors. The opinions in this article reflect Gartmore’s views as at 15th October 2007 and are subject to change.