Issue: June 2005

Baffled by benchmark buzz?
Choose your tools carefully

"Investment speak" is full of talk about benchmarks that sometimes sounds confusing, but in fact their purpose is actually very similar to benchmarks used in real life. When we benchmark ourselves against someone we are setting a standard or point of reference against which we measure our achievements. In its purest sense, we would choose which good points to copy, and which bad points to discard. In this way, we would hope to be even better than the standard or benchmark we have chosen.

The investment world operates along similar lines, although the use of benchmarks is quite varied.

What are benchmarks used for?

Benchmarks are generally used to measure fund performance, manage risk, build fund portfolios and/or set objectives for a fund's returns.

Performance measurement:

The performance of a portfolio is compared against its benchmark and the difference calculated. Analysis is then done to find out where the portfolio manager added to or detracted from performance. This gives insight into how the portfolio was structured, and the various positions that the portfolio manager has taken. To illustrate, the table below compares the median equity returns of the largest asset managers in South Africa (as provided by Riscura) to the returns on the All Share Index since 1998 and calculates the relative out-or under-performance.

Performance measurement (in the form of attribution analysis) would then analyse this out-or underperformance. Typical areas of focus would be: which shares have added to or detracted from performance?; or which sectors have had the best returns and has there been sufficient exposure to them?

Risk management:

Tracking error indicates how likely it is that the portfolio's performance may be different to that of the benchmark. Some portfolios are managed within a maximum tracking error limit, thus setting out in advance how big these differences can be. However, this does not necessarily control the absolute risk of the portfolio - just the risk relative to its benchmark. If the inappropriate benchmark has been chosen in the first place, then a tracking error limit cannot control the overall risk in the portfolio.

Portfolio construction:

Sometimes benchmarks are used to construct an equity portfolio. When deciding how much of any share to have in a portfolio, a portfolio manager may refer to that share's size in the benchmark. The risk with this type of benchmark usage is that the fund manager may be tempted to "hug" the benchmark very closely, instead of investing in those shares that offer the best value.

Return objectives:

Benchmarks are often used to calculate a desired return objective, for example a fund could aim to achieve performance of the All Share Index plus 2%. Inflation is a very important factor in determining a return objective. Many funds will set an objective in relation to inflation, for example a targeted return of inflation plus 5%. Any investment which beats inflation has provided a "real" return. This is essential in planning for retirement in order to ensure that savings maintain their value in an inflationary environment.

Year Ended
Mediun Equity Return
ALSI Return
Out or Under Performance
Dec 1998 -10.8 -10.1 -0.7
Dec 1999 66.9 61.4 5.5
Dec 2000 -0.4 -0.1 -0.3
Dec 2001 21.4 29.0 -7.6
Dec 2002 0.5 -8.2 8.7
Dec 2003 24.3 16.1 8.2
Dec 2004 40.5 25.4 15.1
Annualised Returns Dec 1998 - Dec 2004 17.9 13.9 4.0

Types of benchmarks

Benchmark choice is pretty wide in South Africa, but some of the more commonly used benchmarks for equity investments are:

The All Share Index or ALSI:

This consists of all shares listed on the Johannesburg Stock Exchange. Each share is weighted according to its size, with some of the larger shares exceeding 10% of the index. This can pose some practical challenges as many investment vehicles are prohibited by law from holding more than 10% of their portfolios in any one share. How can portfolio managers aim to outperform a benchmark when they are restricted from owning the full contents of that benchmark?


In an attempt to address the size issues of the ALSI, the CAPI was introduced in July 2003. This index limits all stocks to a maximum of 10%, thus "capping" them.


This index adjusts for the effect of foreign ownership in our market. Certain shares are listed both in South Africa and offshore, known as dual-listed shares. The proportion of these dual-listed shares held by foreigners is excluded when calculating the size of these shares in the index. This down weights more shares than the CAPI does and thus further reduces the size issue of the ALSI.

So which benchmark is best?

Sadly, there is no one-size-fits-all benchmark. The choice of benchmark has to be made in relation to the investment profile and risk of the underlying fund. Clear objectives have to be set and used as a framework for choosing the best benchmark.

Creating a benchmark that is easy to beat is not necessarily in the investor's best interests. What is in the investor's best interests is ensuring that the benchmark will assist in delivering real risk-adjusted returns over the longer term. Ultimately benchmarks should be another tool to ensure that retirement savings do actually finance our retirement.