“Past performance is not a guide to future returns” appears on almost every investment-related disclaimer, yet despite this clear warning, investment selection decisions often remain routinely anchored to performance figures.
For advisers and clients, the relevance this data matters...
This article explains why relying on three-year performance as a core metric for investment selection decisions can create extremely poor short term investment returns – and why a disciplined review process leads to more consistent outcomes.
Three-year returns are still among the most common reference points in fund selection, portfolio reviews, and client conversations – despite being one of the least reliable indicators of what might come after. The issue isn’t that performance data is misleading, but that short performance snapshots are unstable and highly context-dependent.
Three years feels sensible, recent enough to feel relevant, long enough to appear meaningful, and easy to compare, yet markets don’t move in neat three-year chapters.
When decisions are made on that basis, outcomes are far less predictable than the league tables suggest. For us, the discipline of review and process matters far more than any single snapshot of past performance.
Why three-year performance feels so compelling
For advisers and investors alike, past performance figures create a comforting narrative: ‘this has worked’. It appears to demonstrate skill, so why wouldn’t it continue?
That confidence, however, rests on a critically misplaced assumption: that performance persists.
In practice, that assumption rarely holds.
What actually happens after three years?
When funds in the IA Global sector are grouped and selected purely based on rolling three-year returns, performance leadership breaks down far more quickly than most people expect. On average, a fund chosen from the top half of the three-year performance table remains there for only around twelve months before falling into the lower half, the observation below being taken when reviewing all episodes within the 5 year period ending November 2025.
The chart below indicates the performance of the top half of the IA Global sector based on 3 year returns as as the categorisation metric with a static start point at the end of 2020.
If you narrow the selection criteria to the top 30% of positive three year performance, and that persistence drops to roughly ten months. Focus on only the top 10%, and it falls again to around six months of relevance before decaying.
In other words, the more impressive the three-year ranking appears, the quicker (on average) that performance subsequently subsides.
That shouldn’t be surprising, as we all instinctively understand that extreme statistical outcomes are more likely to mean-revert when looking at any dataset – but fund managers often claim to be able to defy gravity. Some can, most don’t.
The central weakness of relying on three-year snapshots is that they show only what has recently gone right, rather than highlighting what is likely to persist from a behavioural perspective.
Context matters: understanding where we are in a cycle
Looking at long-term market history, according to MSCI data between 1946 and 2024, significant drawdowns occur roughly every 2 to 3 years. This means that a typical three-year performance window is more likely than not to include at least one market fall – and often the recovery that follows.
A period that begins after a sell-off and ends near a market high can make almost any strategy look skilful by rewarding disproportionate thematic concentration, while the reverse can make a sound, but more diverse approach appear flawed.
Frequent optimisation – a simple fix?
One response to the instability of strong performance is to review and refresh holdings frequently – continually replacing underperformers and rebuying those with strong recent returns.
In theory, this can maintain performance category adherence. In practice, it introduces new challenges.
The window of outperformance is narrow, outcomes become highly dependent on dominant market themes, and portfolio turnover increases materially.
Even where performance is maintained, the approach relies on constant portfolio optimisation, making it impractical for most investors both operationally and behaviourally. To achieve the results below (keeping exposure to the top 20% of funds consistently) investors need only suffer an annual portfolio turnover of c.700%… tempting, but not really viable!
The practical middle ground: Review and rebalance at sensible (consistent) intervals
The most consistent (and practical) outcomes emerge when performance selection is paired with a disciplined review and rebalance process.
Introducing a simple review cycle, anywhere from a 3 to 12 month basis, materially improves performance persistence, but only when extremes are avoided when making the initial selections.
Rather than chasing the very top of the performance table, the most consistent results come from selecting funds in the lower top quartile or upper second quartile – in this example, funds that exist within the top 20-40% of performers.
The example below assumes a quarterly review and refresh of all positions based on rolling 3-year performance data only – as you can see the performance cluster is easily maintained ahead of the central ‘neutral’ grouping of 40-60%,
The critical point to absorb here, is not about whether any single fund is “right” or “wrong”, but that the ongoing process surrounding selection and monitoring matters far more than any one snapshot itself.
In this example, success depends less on identifying winners and more on maintaining disciplined process to filter for appropriate levels of deviation from the mean.
What time horizons reveal about performance
Looking at performance across different time horizons helps explain why short-term metrics can be misleading. Over shorter periods, outcomes are dominated by noise, timing, and market regime, producing extreme results even in diversified portfolios. Three-year figures often appear calmer, but remain highly sensitive to the start and end points, which is why leadership turnover is so quick.
With longer time frames, individual events matter less, and asset allocation plays a greater role.
The range of outcomes narrows not because markets stop being volatile, but because volatility has time to work through the system. Longer horizons don’t remove risk – they reveal it more honestly, and highlight why process matters more than any single isolated performance window.
Investors experience sequences, not averages
Performance tables rarely reflect how investors actually experience markets. Investors don’t live through average returns – they live through sequences of gains and losses, shaped by timing, volatility, and starting points.
Early losses feel very different from later ones. Extended drawdowns test patience, and volatility invites second-guessing. Once doubt creeps in, even a sound strategy can begin to unravel. This is why timing decisions are so damaging in practice.
They are rarely analytical, often emotional, and once investors start stepping in and out, outcomes deteriorate quickly.
What three-year performance is – and isn’t – suitable for
None of this suggests that three-year performance data should be ignored. Used appropriately, it can provide context, help identify extremes, and prompt helpful questions. What it cannot do, on its own, is reliably predict what comes next. The danger lies not in the data itself, but in treating a short performance window as a definitive answer rather than a partial view.
A more honest way to think about performance
Performance persistence is rarer than many investors expect, particularly at the extremes. In practice, outcomes are shaped less by identifying the “best” recent performers and more by how portfolios are constructed, reviewed, and maintained over time.
The most useful shift, therefore, is a simple one of framing;
Instead of asking:
“What’s done best over the last X-period of time?”
A better question might be:
“How do I reduce my reliance on snapshot-based investment selection?”
…further improved by asking…
“How can I identify consistent behaviours that highlight a managers ability to cope with cycle and regime change?”
Incidentally – we can probably help answer either question!
This content is intended for financial professionals only. These are the author’s views at the time of writing and may be subject to change. This content is not intended to provide the basis for any investment advice or recommendation. Any forecasts, figures, opinions, tools, strategies, data, or investment techniques are included for information purposes only.
The information presented is considered to be accurate at the time of production and has been obtained from or based upon sources believed by the author to be reliable and accurate, but no warranty of accuracy is given and no liability in respect of any error or omission is accepted. Please visit our Regulatory Information and Terms of Use pages for more information.
Production