What a Fund Star Rating Can and Cannot Tell You

What a rating actually is

A star rating is a ranking. A fund is placed against others in its category over some past period, usually with an adjustment for how much it moved about along the way, and the resulting position is expressed as a number of stars.

Three things follow immediately from that definition, and most of the confusion about ratings comes from not having noticed them.

It is backward-looking by construction. There is no forecast inside it. Every input already happened, which means a rating is a summary of history presented in the grammar of a recommendation.

It is relative, not absolute. A fund can gain stars while losing money, if it lost less than its category. It can lose stars while making money. The rating answers "compared with these others" and cannot answer "is this any good for me".

It moves. As the past period rolls forward, funds enter and leave the top of the ranking without anything about them changing. A fund can be rerated because a stretch of history dropped out of the window, not because it did anything.

None of this is hidden. It is simply not what a symbol conveys.

What the rating is genuinely good for

It is worth being clear that these are useful instruments when read as what they are.

A rating is a reasonable filter for gross unsuitability. A fund persistently at the bottom of its category over a long window, on a measure that accounts for how much it moved about, is telling you something worth knowing — most often that its costs are high or that its mandate is being executed badly.

It is also a compression of work you would otherwise do badly or not at all. Somebody has computed a risk-adjusted comparison against a defined peer group across a defined window. That is more than most people would do unaided, and it is more than a raw return figure.

And it puts funds into categories, which is quietly the most useful part. Knowing which peer group a fund has been placed in tells you roughly what it is obliged to hold, and that determines more of its future behaviour than its rating does.

The four limits that matter

It is a rating of a period that ended. Whether a good past predicts a good future is a question with an answer, and the answer is not in the rating. That question is taken apart in why past performance is not enough to choose a fund.

The category defines the answer. Change the peer group and the stars change. A fund at the top of a narrow category may be an ordinary performer in a wider one, and the comparison a rating uses was chosen by the rater rather than by you. If your actual alternative is a low-cost index fund rather than the other members of that category, the rating is not comparing against your alternative.

It cannot distinguish skill from mandate. A fund whose mandate happens to suit the period will rate well for the whole period, and rate badly for the next one, with the same people making the same decisions throughout. The distinction requires evidence a rating does not contain.

It is not advice about you. A rating knows nothing about when you need the money, what else you own, or how much of a fall you can sit through. Two people should reach different decisions about the same five-star fund, and the rating cannot help with the part where they differ.

The test that would settle the useful question

The question worth answering is narrow: do highly rated funds go on to outperform their categories? Not whether they did well in the past — that is what earned the rating — but whether the rating has predictive content.

Doing it properly means taking the ratings as they stood on a date, following those funds forward over a later period, and comparing what each rating band delivered. It means including every fund that was rated at the start, particularly those that closed or were merged during the follow-up period, because those disproportionately did badly and dropping them makes every rating band look better than it was. It means repeating it over several separate periods, and fixing the measure of success before looking.

The honest gap

That test needs the rating history for Indian funds, matched to the subsequent return history of every fund that carried a rating, including the ones that no longer exist.

This site does not have that data. We hold index and inflation history and nothing at the fund level. So this page makes no claim about whether Indian star ratings predict anything, in either direction. It would be easy to assert that they do not — the assertion is fashionable and would cost us nothing — and we have not run the test, so we do not make it.

What can be said without any data is everything in the sections above: what a rating is computed from, and therefore what kinds of question it is structurally incapable of answering.

How to use one anyway

Treat a rating as one input with a known shape rather than as a verdict.

Use it to narrow, not to choose. Then decide between what remains on the things that are knowable in advance — cost, mandate, and how much your decision depends on one person continuing to be right. Cost in particular is the input most likely to persist and is subtracted whether the fund performs or not.

Be most sceptical of a rating that has recently improved. That is the case where the window has most likely shifted rather than the fund, and it is also the moment the rating is most likely to be put in front of you.

And if you find yourself changing a holding because its stars changed, check what actually moved. A fund that has been rerated because a market cycle rolled out of a five-year window has not given you new information about itself — the general version of that trap is in how to respond to investment underperformance.

What to take away

A star rating is a compressed, backward-looking, relative comparison against a peer group somebody else chose. Read that way it is a decent filter and a reasonable starting point.

Read as a forecast, it is being asked to do something it was never built to do — and whether it could do it anyway is a question that needs data this site does not hold and has not pretended to.