Open any fund comparison site and the default sort is one-year return. Every “best mutual funds for 2026” listicle ranks by past performance. It is the single most common way Indian investors choose a fund.
This template exists to let you test whether it works, and it is included in the catalog as a hypothesis to falsify rather than a strategy to deploy. That is an unusual thing for a product to ship, so it is worth explaining why.
What the persistence evidence says
The question has a precise name in the literature — performance persistence — and it has been studied for forty years.
S&P’s SPIVA India Persistence Scorecards track what happens to Indian funds sorted into quartiles by performance. The consistent finding across editions is that only a small minority of top-quartile equity funds remain top-quartile over subsequent three-year periods — rates broadly in line with what random reshuffling would produce. Persistence at the top is weak; the more reliable persistence is at the bottom, where poor funds tend to stay poor or close, largely because high costs are persistent in a way that skill is not.
Mark Carhart’s 1997 paper is the classic treatment and reaches a sharper conclusion. He found that most of the apparent persistence in US mutual fund returns disappears once you control for the momentum factor in the underlying stocks. Funds that outperformed were mostly holding stocks that had momentum, not exercising skill; the following year they held the same stocks and momentum did whatever momentum did. The persistence that survived all controls was concentrated in the worst funds, driven by expenses.
The implication is subtle and important: if fund momentum works at all, it may work because you are getting stock momentum in a wrapper — with an extra layer of fees on top. You could get the same exposure more cheaply and with more control by running a stock momentum screen directly.
Why people do it anyway
Three well-documented biases point the same direction.
Extrapolation. Investors project recent returns forward. A fund up 45% last year feels like a fund that returns 45%, even though the sensible prior is that it holds something that had an exceptional year.
Availability. Recent, vivid performance dominates attention. The fund that had a great year is the one in the news, in the ads, and in the conversation.
Recency in a noisy series. One year of fund returns is a tiny sample. The dispersion between the best and worst equity fund in a year is enormous and mostly reflects style and sector positioning rather than skill — a smallcap fund in a smallcap year is not a good fund, it is a fund with a mandate that suited the year.
Sorting by one-year return therefore mostly sorts by “which style just worked,” and buying it is buying the style after it has worked, which is the wrong end of a mean-reverting cycle as often as it is the right end of a trending one.
The costs the backtest will not show you
Two frictions fall entirely on a rotation strategy and neither appears in a NAV series.
Exit loads. Most Indian equity funds charge an exit load — commonly around 1% — on units redeemed within a year of purchase. A monthly rotation strategy redeems almost everything well inside that window, so it pays this on nearly every exit. Against a strategy whose whole claimed edge might be a few percent a year, this is not a rounding error.
Short-term capital gains tax. Equity mutual fund units held under twelve months are taxed as short-term capital gains. A rotation strategy realises gains constantly and converts what would have been deferred, compounding, untaxed appreciation into an annual tax bill. A buy-and-hold investor in the same fund defers that entirely.
Together these mean the rotation strategy must beat buy-and-hold by a meaningful margin before costs merely to match it after. When you look at the equity curve this template produces, mentally subtract both. Neither is in the NAV.
How mutual fund backtesting works here
A few mechanics that make the test possible.
NAV is the price. Every scheme is a security with a daily NAV, so the same close field that carries a stock’s closing price carries a fund’s NAV. Momentum, moving averages and z-scores all work on it unchanged.
Fill timing matches how funds actually work. The engine decides on the previous day’s data and fills at the next day’s price. For a mutual fund that is not an approximation — it is exactly the real process. Place an order before the cut-off and you are allotted units at that day’s NAV, computed after the market closes. The backtest’s timing convention and the fund industry’s allotment mechanics happen to agree.
Fractional units are allowed. You buy rupees of a fund, not whole units, and allotments run to three decimal places.
NAV is net of the expense ratio. The scheme’s costs are already deducted before the NAV is struck, so a NAV-based backtest has the fund’s fees baked in. What it does not include is your exit load, your tax, or the difference between a regular and a direct plan.
The comparison that answers the question
The template is only useful next to a control. Run it, then run the same window with a plain Nifty 50 benchmark, and compare.
If the rotation beats the index by less than the exit load plus tax drag, it has lost. If it beats it comfortably, the follow-up question is whether the excess survives running the same test over a different window — because a strategy that works in one fifteen-year sample and not another has told you about the sample.
A second, more diagnostic test: compare it against the stock momentum template over the same period. If fund momentum is really stock momentum in a wrapper, as Carhart’s result implies, the stock version should do at least as well without the wrapper’s fees.
Try it
The Mutual Fund Momentum Rotation template ranks schemes on twelve-month NAV growth, requires a full twelve months of history, and holds the top ten equally weighted with a monthly rebalance.
The honest framing: this is a tool for answering a question you probably have. If the answer comes back weak — and the published persistence evidence suggests it will — the useful response is not to abandon funds but to consider the low-cost core template, which screens on the one scheme attribute that is known in advance and does not decay.
Further reading
Expense ratio is the only fund number you know in advance is the direct counterpart to this post: the one fund attribute that does predict. Momentum investing in India covers where the same signal has held up, and why the construction has to differ.
Two terms are worth pinning down before reading any persistence table: base rate and benchmark. Momentum itself is defined there too.