Executive Summary
This study examines how often a lump-sum investment in Direct Growth Flexi Cap schemes ended a rolling holding period with a negative annualised return. The analysis uses internally validated daily NAV data and tests 5 horizons: 1, 3, 5, 7 and 10 years.
Key findings in the observed windows:
- Loss probability was 17.29% at 1 year (113,077 daily rolling windows across 40 Direct Growth schemes), 1.97% at 3 years, 0.29% at 5 years, and 0.0% at both 7 years and 10 years.
- The zero-loss horizon, defined as the first tested horizon with no loss window, is 7 years (45,828 daily rolling windows across 23 Direct Growth schemes).
- The worst 7-year outcome was +3.84% annualised (Taurus Flexi Cap Fund, 2013-03-23 to 2020-03-23).
- At 5 years the worst outcome was -4.91%, and only 2015 start-year windows showed any loss (2.8%).
All statistics are lump-sum, point-to-point annualised returns. Other holding periods were not tested, so 7 years is the first zero-loss horizon among those tested, not a proven boundary.
Methodology
| Parameter | Specification |
|---|---|
| Source | nav_history daily NAV |
| Plan selection | Direct plan, Growth option, 1 series per scheme |
| Universe | Currently active schemes only (survivorship bias: closed/merged schemes excluded) |
| Excluded schemes | 6 (no Direct Growth plan) |
| Data floor | 2013-01-01 (Direct plans launched Jan 2013) |
| Window step | Daily (calendar), NAV forward-filled over non-trading days |
| Return metric | Annualised CAGR per window |
| Dropped windows (gap) | 0 |
| Category label | Reflects current SEBI classification; earlier windows use today's membership |
| Category since | 2020-11-01 |
Each horizon's window count is the number of daily rolling windows pooled across the stated number of Direct Growth schemes. Window start dates run from 2013-01-01 to a horizon-specific end date, so that every window is complete. Scheme counts differ by horizon because a scheme needs sufficient history to contribute a complete window. Daily windows overlap heavily, so they are not independent observations.
Loss Probability by Holding Horizon
| Horizon | Daily rolling windows | Direct Growth schemes | Loss probability | Last window start |
|---|---|---|---|---|
| 1 year | 113,077 | 40 | 17.29% | 2025-10-01 |
| 3 years | 85,381 | 36 | 1.97% | 2023-10-01 |
| 5 years | 63,257 | 27 | 0.29% | 2021-10-01 |
| 7 years | 45,828 | 23 | 0.0% | 2019-10-01 |
| 10 years | 23,208 | 18 | 0.0% | 2016-10-01 |
Loss probability declined at each successive tested horizon until it reached 0.0% at 7 years. The 10-year horizon also recorded 0.0%. The zero-loss horizon is therefore 7 years in this snapshot.
Return Distribution by Horizon
| Horizon | 5th percentile | Median | 95th percentile | Std. deviation | Worst window | Best window |
|---|---|---|---|---|---|---|
| 1 year | -7.4% | 12.1% | 58.93% | 20.53 | -38.24% | 145.54% |
| 3 years | 4.29% | 16.19% | 28.0% | 7.02 | -10.48% | 53.16% |
| 5 years | 5.88% | 15.53% | 25.02% | 5.83 | -4.91% | 41.96% |
| 7 years | 9.35% | 14.86% | 20.52% | 3.26 | 3.84% | 27.11% |
| 10 years | 10.38% | 15.38% | 20.21% | 2.94 | 7.3% | 26.72% |
The spread between the 5th and 95th percentiles narrows as the horizon lengthens, and the reported standard deviation falls from 20.53 at 1 year to 2.94 at 10 years. The worst 1-year window (HDFC Flexi Cap Fund, 2019-03-23 to 2020-03-23) returned -38.24%. The worst 3-year window (Taurus Flexi Cap Fund, 2017-03-23 to 2020-03-23) returned -10.48%. The worst 5-year window (Taurus Flexi Cap Fund, 2015-03-23 to 2020-03-23) returned -4.91%. The worst 7-year window (Taurus Flexi Cap Fund, 2013-03-23 to 2020-03-23) returned +3.84%. The worst 10-year window (LIC MF Flexi Cap Fund, 2015-03-03 to 2025-03-03) returned +7.3%.
The best 7-year window was +27.11% annualised (quant Flexi Cap Fund, 2019-08-25 to 2026-08-25).
Loss Probability by Start Year
Aggregate loss probabilities can conceal concentration. The table below reports loss probability by window start year. A dash indicates that no complete window exists for that start year at that horizon.
| Start year | 1-year loss probability | 3-year loss probability | 5-year loss probability |
|---|---|---|---|
| 2013 | 4.36% | 0.0% | 0.0% |
| 2014 | 8.08% | 0.0% | 0.0% |
| 2015 | 39.75% | 0.0% | 2.8% |
| 2016 | 0.14% | 0.0% | 0.0% |
| 2017 | 20.67% | 24.01% | 0.0% |
| 2018 | 35.06% | 0.05% | 0.0% |
| 2019 | 43.41% | 0.0% | 0.0% |
| 2020 | 0.0% | 0.0% | 0.0% |
| 2021 | 22.87% | 0.0% | 0.0% |
| 2022 | 13.21% | 0.0% | - |
| 2023 | 0.0% | 0.0% | - |
| 2024 | 20.89% | - | - |
| 2025 | 19.53% | - | - |
Observations from the table:
- 1-year loss probability varied widely by start year, from 0.0% (2020 and 2023 start years) to 43.41% (2019 start year).
- At 3 years, the 2017 start year accounted for the highest start-year loss probability (24.01%), with 0.05% for 2018 and 0.0% for other start years.
- At 5 years, only 2015 start-year windows showed any loss (2.8%).
- At 7 years, every start year from 2013 to 2019 recorded 0.0%. At 10 years, every start year from 2013 to 2016 recorded 0.0%.
The data show where losses occurred; they cannot establish why. No causal attribution is made.
Inflation Reference
Using the RBI CPI upper tolerance band of 6.0% as the threshold, the share of windows with annualised returns above that threshold was 65.82% at 1 year, 92.89% at 3 years, 94.81% at 5 years, 99.76% at 7 years and 100.0% at 10 years. A window can therefore avoid a nominal loss and still fall short of this threshold, which is a distinct question from loss probability.
Limitations
- Survivorship bias: only currently active schemes are included; closed or merged schemes are excluded. Loss probabilities for the full historical fund population may differ.
- Overlapping windows: daily windows within a horizon share most of their underlying NAV path. The window counts describe sample size, not independent trials.
- Shrinking and shifting samples: longer horizons have fewer schemes (23 at 7 years, 18 at 10 years) and window start dates that end earlier (2019-10-01 and 2016-10-01 respectively). Longer-horizon results therefore reflect a narrower range of start dates.
- Category membership: earlier windows use today's category membership, and the category has existed since 2020-11-01.
- Limited horizons: other holding periods were not tested.
- Pre-tax, lump-sum basis: returns are NAV-based annualised figures for a single lump-sum entry and exit. Per the snapshot's regulatory facts, long-term capital gains on equity mutual funds held over 1 year are taxed at 12.5% on gains above ₹1.25 lakh per financial year, and short-term gains on holdings up to 1 year are taxed at 20% (Finance (No. 2) Act 2024, as of 2026-10-01). These effects are not reflected in the reported returns.
- Past distributions: historical distributions do not predict future results, and a 0.0% observed loss rate is not a statement about future outcomes.
Data Integrity Statement
The figures in this paper derive from an internally validated daily NAV dataset covering Direct plan, Growth option series for currently active Flexi Cap schemes, from the data floor of 2013-01-01. 0 windows were dropped for data gaps. 6 schemes were excluded for lacking a Direct Growth plan. All values are reported as computed in the quantitative snapshot generated on 2026-10-04, without derived or re-estimated statistics.
This is educational content, not investment advice; consult a SEBI-registered investment adviser.