Finding the top performers

The PortfolioConstruction Forum Managed Funds Outperformance Report aims to identify funds that have outperformed their peers consistently over time. This extract presents analysis of wholesale Australian-domiciled international equity funds in the Mercer IS survey (148 funds in total). Returns are reported on an after-tax and after-management fee basis.

Use of rolling average returns

Rolling average returns are more representative of a fund’s return history than returns over a single time period. It’s not unknown for fund managers to cherry-pick the best returns to a recent month end when promoting a fund.

Rolling average returns smooth out aberrations in a fund’s returns over time and reveal what has been typical over time. For example, a three-year-old fund has many three-month returns in its history — after the first three months there is a new three-month return at every month end, giving 34 three-month returns over the three-year period.

A rolling average three-month return averages all 34 three-month returns over that period, rather than looking at just the most recent three-month return, which could be an aberration and very different to the normal returns of the fund.

Similarly, a rolling one-year return is the average of all the one-year returns of a fund over the period examined (for example, 25 one-year returns over a three-year period).

Decile rankings

Decile rankings help to show the extent a fund outperforms or underperforms its peers. Decile rankings are similar to quartile rankings but give a finer grading. Recent academic research has also shown that decile rankings are somewhat consistent over time, particularly at the very top and very bottom deciles.

For example, if the average one-year returns of a group of funds range from —25 per cent to 25 per cent, then the range is 50 per cent, giving 10 deciles of 5 per cent (50/10=5). Each fund is assigned to a decile based on the fund’s return.

In our example, a fund with a return greater than 45 per cent is in the first (top) decile. A fund achieving better than 40 per cent and up to 45 per cent is in the second decile, and so on. It follows that some deciles may have no funds in them (no fund has achieved a return in that decile range), while other deciles may have many funds in them.

Time frames

The analysis splits funds in the universe into two subsets: those for which there is data for at least three years (that is, funds that are at least three years old), and those for which there is data for at least five years (that is, funds that are at least five years old).

For each subset, results over three time frames are considered:

~ average one-month, rolling average three-month, and rolling average one-year returns over the fund’s life to December 31, 2005;

~ average one-month, rolling average three-month, and rolling average one-year returns over the three years to December 31, 2005; and

~ average one month, rolling average three-month, and rolling average one-year returns over the five years to December 31, 2005.

The three-year and five-year analyses are an important contrast to the ‘since inception’ analysis, as the latter can contain returns that were achieved under very different investment staff, processes and market environments and therefore ‘since inception’ results can differ markedly from results over the past three and five-year periods.

It is important to note that the analysis includes all funds that reported data in the time period, whether they reported for the whole period or not (thus removing survivorship bias), but for space reasons those no longer reporting at December 31, 2005, are not shown on the graphs and table if they were not in the top two or bottom two deciles.

Picking winners

There is no easy way to identify fund managers that will outperform in the future. Quantitative analysis can highlight funds that have consistently achieved higher returns than their peers in the past. Academic research has confirmed that there is some persistency to top decile and bottom decile rankings over time.

However, PortfolioConstruction Forum considers qualitative research a necessary complement to quantitative analyses when considering how a fund’s returns were achieved, and whether the processes, resources, philosophies, and disciplines involved in achieving the return are likely to be replicable in the future. Of course, much depends on the skills and experience of the qualitative researcher and the validity of the process the qualitative research used.

Deidre Keown is an analyst with PortfolioConstruction Forum.