Contents
Rolling Returns are a series of overlapping period returns computed by sliding the measurement window forward by a fixed interval (daily, weekly, or monthly). Instead of a single point-to-point return, rolling returns produce a distribution of returns across all possible start dates in a historical dataset — eliminating survivorship bias from cherry-picked windows.
Why Rolling Returns Beat Trailing CAGR
A trailing 5-year CAGR as of today measures one specific 5-year window. An investor who started 5 years and 3 months ago, or 4 years and 9 months ago, has a different experience. Rolling returns model all possible investors across all entry points — the result is a realistic range of outcomes, not a single potentially flattering figure.
Example: A flexicap fund's 5-year rolling returns computed monthly from January 2010 through December 2024 (approximately 180 rolling 5-year windows):
- Median 5-yr CAGR: 14.2%
- Minimum 5-yr CAGR: 4.1% (window starting near 2007 peak)
- Maximum 5-yr CAGR: 22.8% (window starting near March 2020)
- % of windows delivering >12%: 78%
A single trailing 5-yr CAGR from a COVID-trough start would show 22%+ and be misleading. The rolling distribution reveals that 22% of windows underperformed 12%.
Rolling Return Calculation — Mechanics
For a 3-year rolling return using monthly data:
- Take the NAV/price series from January 2010 to December 2024 (monthly, 180 data points)
- Window 1: Jan 2010 → Dec 2012 (36-month CAGR)
- Window 2: Feb 2010 → Jan 2013 (36-month CAGR, shifted one month)
- Continue until last window: Jan 2022 → Dec 2024
- Result: ~144 overlapping 3-year CAGRs, each representing a different investment cohort
The distribution (mean, median, percentiles, % beat benchmark) is far more informative than a single CAGR.
Rolling Returns and Benchmark Comparison
For evaluating active fund managers, plot the fund's rolling returns alongside the benchmark TRI (Total Return Index) rolling returns for the same windows. Count the % of windows where the fund outperformed the benchmark (hit rate). A quality active fund should demonstrate:
- Hit rate >60% across 3-yr and 5-yr rolling windows
- Outperformance margin in winning windows > underperformance margin in losing windows
- Consistency across different market regimes (bull, bear, sideways)
SEBI and AMFI Disclosure Framework
SEBI does not specifically mandate rolling return disclosure in scheme factsheets — trailing CAGR remains the regulatory minimum per SEBI MF Regulations 1996, Schedule VII and SEBI Circular SEBI/HO/IMD/DF2/CIR/P/2021/647. However, AMFI's investment education content and most AMC research publications (HDFC AMC, Nippon, Mirae, SBI MF) voluntarily publish rolling return analysis. Value Research and Morningstar India also provide rolling return data as standard analytics.
SEBI's Stewardship Code for mutual funds (SEBI Circular SEBI/HO/IMD/DF2/CIR/P/2020/228) encourages AMCs to engage constructively with portfolio companies and evaluate manager performance over full market cycles — rolling returns are the natural analytical tool for this evaluation.
Source: SEBI MF Regulations 1996, Schedule VII; SEBI Circular SEBI/HO/IMD/DF2/CIR/P/2021/647; SEBI Stewardship Code SEBI/HO/IMD/DF2/CIR/P/2020/228; AMFI investor education materials (amfiindia.com).
SIP Investors and Rolling Returns
SIP investors experience a return that approximates the average rolling return across their SIP tenure — each instalment effectively starts a new investment window. If a fund's 5-year rolling return median is 12% with 80% of windows delivering >10%, a SIP investor's XIRR over 5 years will likely be in that range. Rolling returns are thus the conceptual bridge between point-to-point CAGR and SIP XIRR.
Comparison Table
| Metric | What It Shows | Best Use |
|---|---|---|
| CAGR | Single window, point-to-point | Lump-sum comparison; factsheet trailing returns |
| XIRR | Personal rate of return on actual cash flows | SIP/SWP portfolios |
| Rolling Returns | Distribution of all possible windows | Manager consistency; SIP suitability analysis |
Related terms: CAGR, XIRR, Alpha, Sharpe Ratio.
Past performance is not indicative of future returns. ARN-314872. APMI APRN-01658. Content is informational.