A rolling return table is usually read as evidence that holding longer earns more. Run it on Indian equity and it says something narrower and more useful: holding longer does not raise the average outcome, it raises the worst one.
A trailing “5-year return” is one number produced by exactly one start date and one end date. A rolling return recomputes the same window from every available start date and reports the spread. What follows runs that calculation on the Nifty 50 across 1991-2026 and separates two claims that are routinely merged.
The Spread Collapses Toward a Stable Centre
Nifty 50 rolling CAGR by holding period, 1991-2026. Minimum, maximum and average annualised return across every rolling window of each length; the 35-year figure is a single fully elapsed window, not a range. Source: BSE/NSE historical data via BMS Money.
- 1-year holds: ranged from -47.1% to +134.1%, averaging 17.4% - 9 of 35 rolling years were negative.
- 5-year holds: -6.2% to +37.5%, averaging 11.2% - only 1 of 31 rolling windows was negative.
- 7-year holds: 0.0% to +24.5% - zero of 29 rolling windows were negative.
- 10-year holds: never fell below +2.5%, averaging 11.3%.
The Average Stops Moving After Year Three
The average column runs 17.4% at one year, then 12.2%, 11.2%, 11.2%, 11.3%, 11.8% - essentially flat from year three onward. It does not climb with holding period.
The one-year average is in fact the highest number in the table, inflated by outsized single windows - one at +134.1% - that dilute the moment they are compounded into anything longer. An investor reading the table as “longer holding, higher return” is reading a column that does not support the claim.
What Moves Is the Floor
The minimum rises from -47.1% at one year, to -6.2% at five, to a floor that never dips below zero from year seven onward. That is the entire effect, and it is a large one.
Stated precisely: holding longer has not made this index return more on average - it has made the worst plausible outcome dramatically less bad. That is a claim about dispersion, not about expected return, and it is the mechanism a rolling table is built to expose. It also explains why the effect matters most to investors with a fixed liability date, for whom the floor is the binding constraint rather than the mean.
One Market, One Sample
Nothing here guarantees the next thirty-five years resemble the last thirty-five. A single index’s history remains one sample however long it runs, and India’s covers a period of structural growth that may not repeat.
What the table does establish is methodological rather than predictive: a single trailing CAGR quoted as of one date cannot show dispersion at all, because dispersion is exactly what a single window averages away.
What This Means for Allocators
Quote the range, not the point. A trailing CAGR answers what happened to one hypothetical investor with one start date, and the reader is rarely that investor.
Match the holding period to the floor rather than the mean when there is a liability date. The seven-year figure is doing different work from the ten-year average, and only one of them speaks to whether the money will be there.
And read the minimum column first. It is the one that moves with time, and it is the one a single CAGR can never disclose.
Back to Writing