How to review a mutual fund — and the basket behind your goal

"This fund returned 18.2% over five years." It is the sentence every factsheet leads with, and it is the weakest evidence in the document. It describes one journey between two dates that somebody else chose — and it says nothing about what you would have lived through to get there. A real review asks four questions instead. And when the money belongs to a goal, it asks them of the whole basket, not one fund at a time.

Why one trailing return is not a review

A trailing return is a point-to-point measurement: value on one date, value on another, annualised. Both ends are accidents. Shift the start by a few months and the same fund can report a materially different five-year number, because the figure quietly inherits whatever the market happened to be doing on those two days. That is why a fund's headline return can look transformed after a single strong quarter — nothing about the fund changed, only the window.

Worse, it hides everything you actually experience as an investor: how far the fund fell along the way, how long it stayed down, whether its result came from skill or simply from owning a category that was in favour. Those are the things that decide whether you stay invested — and staying invested is most of the return.

1. Return, measured against risk

Return on its own is half a sentence. The other half is what you had to sit through to earn it, and it comes in three parts:

Reading them together changes conclusions. A fund returning 14% a year with a worst fall of −22% and one returning 15% with a worst fall of −41% are not "roughly the same, one slightly better". The second asked far more of you, and if your goal had arrived during that fall, it would have arrived at the bottom.

2. Rolling returns — every start date, not just one

This is the fix for the point-to-point problem, and it is the most useful single view in fund research. Instead of one five-year run, you compute the five-year return from every start date in the fund's history and look at the whole distribution:

• Average and median — the typical five-year outcome, not the lucky one.
• Best and worst run — the range of what five years has actually meant here. The worst rolling run is a far better stress test than the worst single year.
• How often it made money — the share of runs that finished positive. On a long enough horizon, in a broad equity fund, this figure is often more reassuring than the average return is impressive.
• How often it beat its benchmark — the share of runs that finished ahead of the index over the same dates. This is the question people think a trailing return answers.

That last one deserves emphasis, because the weaker version of it is everywhere. "The fund averaged 14% and the index 12%" is perfectly compatible with the fund losing in most individual windows and winning enormously in a few. What you want to know is: of all the start dates I could have picked, in how many would I have been ahead? That requires pairing each of the fund's runs with the index's run over the same window and comparing them one to one.

Two honest-measurement rules go with it. A window labelled "1 year" must contain one-year runs — if a series has a hole in it, a run can silently span far longer and still be printed in the one-year row. And a comparison is only fair if both series cover the same period; a fund with eight years of history has no ten-year rolling returns, and no amount of arithmetic creates them.

3. Comparison with the right benchmark

A fund's return means little until you know what the same money would have done in the plain index for its category. But "compare with the benchmark" hides three traps:

• The wrong index flatters or punishes unfairly. Judge a mid-cap fund against the Nifty 50 and you are mostly measuring the size of the companies it holds, not the work of the manager. Compare like with like — a mid-cap fund against a mid-cap index, a flexi-cap against a broad-market one.
• The index must exist for the whole window. Several Indian mid-, small- and broad-market series are much younger than the funds tracking their segment. Comparing ten years of fund against six years of index reports the four missing years as outperformance — the honest answer there is "not available", not an estimate.
• The index must explain the fund. Beta says how much the fund moved for each 1% the index moved; alpha is the return left over after accounting for that exposure. Both rest on the index actually describing the fund's behaviour, measured by the fit (R-squared). Where the fit is poor, beta and alpha are valid arithmetic and meaningless information — and should be withheld rather than printed.

Get all three right and the comparison starts to separate two very different things: how much of your result came from the asset class, and how much from the fund. Most of it is usually the asset class — which is exactly why asset allocation deserves more of your attention than fund selection does.

4. Upside and downside capture — the shape of the fund

The two capture ratios split behaviour into rising and falling markets, and together they say something no single number does.

Upside capture of 78 means that across the months the index gained, the fund compounded 78% as much as the index did. Downside capture of 79 means that across the months the index fell, the fund took 79% of the fall.

The pair is the point. A fund that captures 78% of the rises and 79% of the falls is essentially tracking its index with a small drag. One that captures 92% of the rises and 65% of the falls trailed in every bull run and gave back much less in every crash — it may well end up ahead over a full cycle, and it will certainly have been easier to hold. The reverse shape — more of the falls than of the rises — is the one worth spotting early, because it tends to show up as a good trailing return in a long bull market and nowhere else.

Two conventions matter if you compute these yourself. They are measured monthly, not daily — daily up and down days are dominated by noise, and the same fund measured both ways gives materially different ratios. And they are compounded across the up months and the down months, not averaged, so the comparison is against what actually happened rather than a mean of percentages that never occurred in that order. A ratio computed from a handful of months is not worth reporting at all; a year's worth of up months and of down months is a sensible minimum for each side.

What a fund review cannot tell you

Everything above is history. The evidence that past outperformance persists is weak, and no combination of these statistics is a forecast. What a review genuinely buys you is a clear picture of the risk you are taking — how far this thing has fallen before, how long it stayed down, how much of the index's fall it tends to absorb, and how much of the result was the asset class rather than the manager. That knowledge is useful whether or not the ranking holds.

Be alert, too, to the things statistics do not capture: a change of mandate or category, a merger, a fund that has grown so large its strategy no longer fits it, or a costs structure that quietly eats the difference you are measuring.

Part two: review the goal's basket, not the funds one by one

Almost nobody owns one fund. You own four or five, chosen at different times for different reasons, and together they fund something — retirement, a child's education, a house. That collection is the thing that either reaches the goal or doesn't. Yet it is almost never the thing that gets reviewed.

Diversification you don't actually have

Five equity funds can feel like five decisions and behave like one. If they hold much the same large-cap stocks, their returns move together, and the basket's volatility and worst fall barely improve on any single member's. The measurement for this is correlation: how alike the funds have behaved. Two funds that move almost in lockstep are one position wearing two names — and the fee is paid twice.

(Correlation describes behaviour, not composition. It is not the same claim as "these two funds hold 60% of the same stocks", and it should not be read as one.)

The numbers that only exist at the basket level

• The mix's own worst fall. Not the worst of the members — the drawdown of the combined thing, which is what your goal actually experiences.
• The mix's volatility and return per unit of risk. The whole reason to hold several funds is that the combination should be steadier than its parts. Either it is, or the diversification is decorative.
• What a monthly SIP into it would have produced. A point-to-point return on a lump sum and a money-weighted return on a monthly SIP answer different questions, and for most goals the SIP is the honest one — it is how the money will actually arrive.
• How the mix compares with a single index, and with an alternative mix. "Does adding a small-cap sleeve to this basket change the return enough to justify what it does to the worst fall?" is a basket question with a testable answer, and it cannot be answered by looking at the small-cap fund's card alone.

Rebalancing changes the answer — and it isn't free

A basket's result depends on how often you put it back to its intended weights. Left alone, the winners grow into a bigger share and the mix drifts into a risk profile you never chose (the drift problem). Rebalanced yearly, it stays near the weights you picked. Any honest comparison of the two has to note that rebalancing is not costless: every rebalance is a redemption, and a redemption can mean an exit load and a capital gain. Any backtest of a mix that assumes frequent, frictionless rebalancing is flattering itself — which is why the "never rebalanced" case is the baseline worth looking at beside it.

Then tie the basket back to the goal

A basket is only reviewable against a purpose. Three things connect them:

• The horizon. A basket that is fine for a goal 15 years away is the wrong basket for the same goal 3 years away. As a goal approaches, reviewing shades into de-risking — the glide path described in goal-based investing.
• The expectation. You committed money to this basket assuming some rate of return. The review question is not "did it beat the index" but "is it tracking to the number my plan was built on" — and if not, whether the plan or the basket should change.
• The sequence. For anything you will withdraw from — retirement above all — the order of returns matters more than their average. That is sequence-of-returns risk, and it is the reason a basket's worst fall matters more as the goal gets closer.

A practical review rhythm

• Once a year is enough for the numbers — or whenever you rebalance. Monthly checking mostly measures noise and tempts you into trading.
• Review the basket first, the funds second. Start with what the goal's mix did as one thing; only then ask which member is responsible for what.
• Use the same window for everything you compare. Most fund-comparison errors are just two things measured over different periods and printed side by side.
• Prefer "not available" to an estimate. If a fund is younger than the window, if the index doesn't reach back that far, or if there are too few observations, the right answer is a blank — not a number that will be read as a fact.
• Act on structure, not on rank. A mandate change, a category drift, a duplicated position or a horizon that has shortened are reasons to change something. One quarter of underperformance is not.

Where to run this review

The InvestApps Investment Tracker does all of it from real NAV history:

• Discover Mutual Funds (free, no account) — search any open scheme AMFI lists and see its returns, volatility, Sharpe, Sortino, worst fall, rolling returns from every start date, beta and alpha against the right index for its category, and its upside/downside capture. It withholds any figure the data cannot honestly support, and compares a fund with its peers — or with any other fund — over the same window.
• Fund groups (free) — weigh several funds into a named mix and measure the mix as one thing: its return, volatility, worst fall, a monthly-SIP backtest, its rolling returns and how often they beat the index, a correlation grid showing how alike its members have behaved, and a direct comparison against another mix.
• Your own portfolio — map real holdings to goals and buckets, so each part of your book reports its own return against the return you expected of it, measured from your actual daily portfolio value rather than averaged from factsheets.

Nothing in the apps is ranked, scored or recommended, and no preset names a fund. InvestApps.in is not a SEBI-registered investment adviser: these screens report what happened, and the decision stays yours.