Anatomy of a legendary fund run — and why it ended

Every few years a fund manager produces a run so good it becomes a story. The returns are real, the process is articulate, and money floods in. Then, usually, it stops working — and the investors who arrived because of the story take the part of the record that follows it.

This guide dissects the shape of those runs. It deliberately does not name anyone, because naming invites you to go and find the current holder of the title, which is precisely the behaviour the pattern punishes. The anatomy generalises; the individual does not.

The five phases

Phase 1 — obscurity (years 1–3). A small fund, an unfashionable process, low assets. The manager can take meaningful positions in mid and small caps because the fund is small enough that the positions are buildable. Performance is decent and nobody notices.

Phase 2 — the run (years 3–7). The style comes into season. If the process is quality-and-growth, a growth cycle arrives; if it is value, a value cycle does. Returns are excellent, the fund tops its category table, and star ratings rise — which is mechanical, because ratings are computed from past returns.

Phase 3 — the narrative (years 5–8). Interviews, profiles, a memorable articulation of the philosophy. Note that the narrative arrives after the returns and is constructed to explain them. It is a description, not a cause.

Phase 4 — the flood (years 6–10). Assets multiply, often ten-fold or more. And here the strategy begins to change without anyone deciding to change it: the positions that generated the return cannot be built at the new size. The portfolio drifts up the market-cap curve, the number of holdings rises, position sizes fall. The fund becomes something else while keeping the same name and manager.

Phase 5 — the reversion (years 8–12). The style cycle turns. Performance lags for years. The investors who arrived in phase 4 — the majority, by rupees — experience only this phase, and most leave near the bottom.

Separating skill from tailwind

This is the analytical core, and it is answerable with numbers already on a fund page.

Was the benchmark the right one? A mid-cap-heavy portfolio measured against a large-cap index produces flattering alpha that is really a cap-size effect. Check R² — if the benchmark explains little of the fund's movement, the alpha figure is not describing skill.

Did it work in more than one regime? A record spanning only a growth cycle, or only a bull market, tells you the style was in season. A record spanning both 2008 and 2013–17 and 2020 tells you something about the process.

What does the information ratio say? It measures outperformance per unit of active risk — the cleanest available separation of skill from either the market's return or the manager's willingness to take risk against the benchmark.

How consistent, on rolling returns? A great trailing number can be one exceptional year inside an otherwise ordinary record. Rolling returns show the distribution and expose that immediately.

Was the drawdown proportionate? A manager who beat the index by taking substantially more risk has not demonstrated skill — they have demonstrated leverage to the same factor.

How much did concentration contribute? In a focused fund, a run can rest on two or three positions. That is a legitimate outcome and a poor basis for forecasting the next decade.

Why runs end

Rarely because the manager got worse. Usually because of one of these:

• The style cycle turned. The most common cause by far, and the least discussed at the time.
• Size destroyed the strategy. The phase-4 problem. This is the one you can actually see coming, by watching AUM alongside the market-cap distribution of the portfolio.
• The manager left. In a concentrated, judgement-driven fund this matters far more than in a process-driven one. See fund manager changes.
• The inefficiency was arbitraged away. Other managers adopted the same approach and the edge compressed.
• It was luck. With thousands of funds, extraordinary records occur by chance alone. Distinguishing a skilled manager from a lucky one requires more years of data than most careers contain.

What to do with all this

1. Do not buy a fund because of a run. By the time the run is visible, you are entering at phase 4 — the stage where the returns compress.
2. Check the size before the record. Ask whether the current AUM permits the process that generated the record.
3. Weight process over personality. A documented, repeatable process survives a manager's departure. A star's judgement does not.
4. Prefer records spanning multiple regimes. Ten years including a bear market beats five spectacular ones.
5. Size any active bet as a satellite. A diversified or index core does not depend on any of this being right.
6. Accept the honest baseline. The evidence on persistence of outperformance is not encouraging — which is why indexing the core is the default this site recommends, and why an active satellite should be a considered exception rather than the foundation.

Pitfalls to avoid

• Chasing the current star. Their best years are typically behind them by the time you have heard of them.
• Mistaking the narrative for the cause. It was written to explain returns that had already happened.
• Ignoring AUM growth. It is the most predictable killer of a good strategy.
• Following a manager to a new fund house. You are buying the person without the team, the analysts or the process that supported them.
• Judging a lagging manager over three years. Style cycles are longer.
• Assuming the record is transferable. A record built in mid caps at ₹1,500 crore says little about large caps at ₹40,000 crore.

Key takeaway

Great fund runs follow a recognisable five-phase arc — obscurity, the run, the narrative, the flood of assets, and reversion — and most investors arrive at phase four, which is when the size of the fund has already begun to change the strategy that produced the record. Separate skill from tailwind using the numbers rather than the story: the right benchmark, R², the information ratio, rolling returns, and whether the drawdown was proportionate to the outperformance. Weight process over personality, watch AUM as closely as return, and hold any manager bet as a satellite around a core that does not depend on it.