Active Return requires Active Risk

In order to justify an active investment management strategy, a portfolio manager is required to demonstrate that she can, over time, generate a return in excess of a predetermined benchmark. The generation of this excess return, or alpha, is only possible if the portfolio is constructed to look different from the benchmark. In general, portfolio managers utilise two methods to differentiate their portfolios from the relevant benchmark and subsequently generate excess returns. These methods are as follows:

1. Ignore the benchmark completely

Under this method, the portfolio manager pays no attention to the composition of the benchmark when constructing the portfolio. To illustrate my point let us examine an equity fund, which would generally be measured against the JSE All Share Index (a market capitalization weighted index). In addition let us make the assumption that the portfolio manager has a bearish outlook on Sasol. This method allows the portfolio manager to totally ignore Sasol’s weighting in the index and have a zero allocation to this stock. For example, even if Sasol constituted 8% of the index, the portfolio manager would be happy to hold none. The risk of constructing a portfolio in this manner is that the portfolio could underperform the All Share Index over short periods of time if Sasol outperforms and drives market returns.

2. Construct the portfolio relative to the benchmark

Under this method, the portfolio construction process is driven by the composition of the benchmark. Once the portfolio manager has assessed the merits of a particular stock, its weighting in the portfolio is determined by referencing the weight in the benchmark. For example, if a manager is bearish on Sasol, she would allocate less than the benchmark weighting to this stock within the portfolio. The extent of the deviation from the benchmark weighting would depend on both the individual merits of the stock and internal limits set by the portfolio manager. It is therefore unlikely for a portfolio manager using this methodology to have a zero allocation to a stock that constitutes a large part of the index (Anglo American, BHP Billiton, MTN etc). This method of portfolio construction can be viewed as a risk mitigating technique as it should ensure that portfolio performance does not stray too far from the benchmark.

Tracking error (also referred to as active risk) is a quantitative measure that can be used to illustrate how closely a portfolio tracks a particular benchmark. For example, an index tracker fund should have a tracking error close to zero as it aims to mimic a particular index, whereas an actively managed fund will have a tracking error greater than zero.

Using 10 years of monthly data (18/06/2000 to 18/06/2010) we can examine to what extent some equity managers have deviated from the benchmark (All Share Index).

 (Click on image to enlarge)

Data Source: Moneymate

Some interesting results are revealed when the annualised return is plotted against the annualised tracking error for various equity funds (see chart above).

Firstly, the two index funds (Stanlib Index Fund and Gryphon All Share Tracker Fund) have registered tracking errors of between 2 and 5, indicating that they do not track the index perfectly. One reason for the higher than expected tracking errors is portfolio costs, which cause returns to be diminished. Due to portfolio costs alone, a tracker fund will never deliver a return that is on par with the index that it is tracking. Another real world difficulty is the management of cash flows, which may cause the portfolio manager to have an allocation to cash within the portfolio from time to time. These cash allocations impact returns and cause further deviation from the index return, resulting in a portfolio that does not perfectly replicate the index.

Secondly, funds that have demonstrated that they can outperform the benchmark considerably have done so by having larger tracking errors than peers. In other words, the portfolio manager generated these returns by constructing a portfolio that was vastly different to the All Share Index.

The table below shows annualised performance and annualised standard deviation of these portfolios for the same period (ranked by performance).

Data Source: Moneymate

During this period the All Share Index generated an annualised return of 17.4% with an annualised standard deviation of 20. Notably, funds with larger tracking errors or active risk have actually managed to deliver alpha at a far lower volatility. This is partly due to the portfolio construction methodology used by the portfolio manager coupled with the fact that the All Share Index is heavily concentrated in resources stocks. Portfolio managers that ignore the benchmark are not required to hold resource stocks and are able to avoid this highly volatile sector when they are not offering value. For example, Investec Value generated the best return over this period, with the highest tracking error, but managed to keep volatility relatively low. Interestingly both index funds (Stanlib and Gryphon) have underperformed the All Share Index by 1.6% and 4% per annum respectively over this period.

When investing, it is paramount to understand how the portfolio manager allocates capital to various asset classes and stocks. It is also important to understand if the portfolio manager is cognisant of the benchmark when making these investment-related decisions. Providing the portfolio manager makes astute investment decisions, the further she deviates from the benchmark, the greater the potential to earn excess returns. Lastly, Index Tracker Funds are not without risk as portfolio costs and cash management practicalities cause these funds to deviate from the index that they are tracking.