Rolling Returns
Rolling returns are a series of return calculations computed over a fixed time window (e.g., 1 year, 3 years, or 5 years) starting from every day (or month) within a historical data set, capturing the distribution of returns an investor would have experienced depending on when they invested.
Rolling returns address a fundamental inadequacy of point-to-point return analysis: when you only look at returns from a fixed start date to a fixed end date, the result is heavily influenced by the specific entry and exit conditions. Rolling returns eliminate this selection bias by computing the same time-window return from every possible starting point within the available data, generating thousands of return observations rather than a single figure.
For example, to compute 3-year rolling returns for a fund with 10 years of daily NAV history, one would calculate the 3-year return starting 1 January 2014, then 2 January 2014, then 3 January 2014, and so on through 1 January 2021 (the last date from which a 3-year period ends within the 10-year window). This generates approximately 1,825 data points for a daily rolling analysis or about 84 data points for a monthly rolling analysis. These data points can then be plotted as a distribution or summarised with statistics (minimum, maximum, average, percentage of positive outcomes).
In Indian mutual fund analysis, rolling returns have become a standard tool for evaluating consistency of outperformance. Value Research, Morningstar India, Advisorkhoj, and PrimeInvestor all offer rolling return analysis tools. When evaluating whether a mid-cap fund has consistently beaten its benchmark, rolling 3-year and 5-year return analysis reveals the percentage of rolling periods in which the fund outperformed — a metric far more informative than a single 5-year trailing return.
Rolling returns reveal several important characteristics of a fund's performance profile: the minimum return (worst case for any investor who invested and redeemed within the specified window), maximum return (best case), median return (typical experience), the proportion of positive return periods (probability of capital preservation over the given horizon), and the frequency and magnitude of benchmark outperformance.
For SIP investors, rolling SIP returns (XIRR computed for SIPs starting on every date within the historical window) are even more relevant than lumpsum rolling returns, since SIP flows are averaged over time and do not depend on a single entry point.
SEBI mandates that AMCs disclose performance data in a standardised format, but rolling returns are not a regulatory requirement — they are a voluntary, analytical enhancement provided by third-party platforms and sophisticated investors.