Sequence-of-returns risk in retirement

Sequence-of-returns risk is the most misunderstood retirement risk. Two retirees can have portfolios that grow at the same long-run average return — say 10% per year — and yet end up with completely different outcomes if the timing of the bad years differs. A crash in year 1 of retirement is fundamentally different from the same crash in year 25.

The mechanism

During accumulation, market drawdowns are an opportunity. SIP contributions buy more units at lower NAVs. Subsequent recovery rebuilds the portfolio. The dollar-cost-average mathematically benefits from volatility.

During decumulation (retirement), the dynamics reverse. Each year you sell some units to fund spending. If those sales happen at depressed NAVs, you sell more units to fund the same expense. The remaining portfolio is smaller and has less ground to recover from. The next year's withdrawal is from an already-shrunken base.

A simple example

Two retirees: A and B. Both start with ₹1 cr. Both withdraw ₹6 lakh per year. Both portfolios produce identical 5-year sequences of -10%, -10%, +5%, +15%, +20% — averaging 4% per year. The difference: A's sequence is reversed (good years first, bad years last) and B has them in the listed order.

Retiree A (good years first): year-end balances after withdrawal:

• Year 1: ₹1 cr × 1.20 - ₹6 L = ₹1.14 cr.
• Year 2: ₹1.14 × 1.15 - ₹6 = ₹1.25 cr.
• Year 3: ₹1.25 × 1.05 - ₹6 = ₹1.25 cr.
• Year 4: ₹1.25 × 0.90 - ₹6 = ₹1.06 cr.
• Year 5: ₹1.06 × 0.90 - ₹6 = ₹89 lakh.

Retiree B (bad years first): year-end balances after withdrawal:

• Year 1: ₹1 cr × 0.90 - ₹6 L = ₹84 lakh.
• Year 2: ₹84 × 0.90 - ₹6 = ₹69.6 lakh.
• Year 3: ₹69.6 × 1.05 - ₹6 = ₹67.1 lakh.
• Year 4: ₹67.1 × 1.15 - ₹6 = ₹71.2 lakh.
• Year 5: ₹71.2 × 1.20 - ₹6 = ₹79.4 lakh.

After 5 years with identical returns and identical withdrawals: A has ₹89 lakh, B has ₹79.4 lakh — a 12% gap created entirely by sequence.

Why early-retirement crashes are catastrophic

The first 5-10 years of retirement determine whether the portfolio survives or fails. A crash in year 1 forces large unit sales at low prices; the recovery has less base to work with. Studies on US retirees show that retiring at the top of a market bull (just before a crash) often produces failure outcomes at standard 4% withdrawal rates, while retiring at the bottom of a bear market (just before a recovery) produces portfolio growth beyond the starting point.

Mitigation strategies

Asset allocation glide path

Reduce equity allocation in the 5 years before and 5 years after retirement. Going from 70% equity at age 55 to 40% equity at age 65 reduces the magnitude of any sequence-of-returns shock during the most vulnerable years.

The bucket strategy

Split the retirement portfolio into buckets by time horizon:

• Bucket 1: 2-3 years of expenses in liquid / short-duration debt. Always withdraw from here.
• Bucket 2: 5-10 years of expenses in conservative hybrid or balanced advantage. Replenishes Bucket 1 during normal markets.
• Bucket 3: long-term growth in equity. Replenishes Bucket 2 during good market years; left alone during bad years.

The discipline: never sell equity in a down year. The Bucket 1 cushion gives time for equity to recover before being touched.

Dynamic withdrawal rules

Instead of fixed 4% withdrawal regardless of market state, adjust based on portfolio performance:

• If portfolio dropped > 20% in the last year: reduce withdrawal by 10% for the next year.
• If portfolio grew > 15%: allow withdrawal increase up to inflation + 2%.
• Floor at 3% withdrawal in any scenario; ceiling at 5%.

"Guardrails" approaches like this allow average higher withdrawal across all years vs static 4% — at the cost of some year-to-year spending variability.

Annuitisation for floor income

Using a portion of the corpus to buy an annuity (NPS or commercial) creates a stable monthly income that does not depend on portfolio performance. The annuitised portion is immune to sequence risk. Many planners suggest annuitising enough to cover essential expenses (60-70% of total expense), leaving market-exposed portfolio for discretionary spending.

Delay retirement by 1-2 years

Working an extra 12-24 months past your initial retirement plan can substantially de-risk the portfolio. You skip the worst of an early-retirement crash, you defer drawing on the portfolio, and you may benefit from market recovery before retiring. Sequence-of-returns sensitivity drops sharply with each additional working year.

The empirical evidence

Research on US retirees retiring in years 1965, 1966, 1968, 1969 (high stagflation years) found portfolio failure rates above 50% even at 4% withdrawal. Retirees in 1982, 1990, 2002, 2009 (post-crash years entering recovery) had near-100% success rates with the same parameters. The macro environment at retirement matters enormously.

Indian markets have similar though less-studied patterns. Investors retiring in early 2008 (just before the GFC) faced sequence risk; those retiring in 2003 (post dot-com / SARS) saw immediate strong returns.

The fundamental insight

For retirees, the average return over 30 years is far less relevant than the order in which the returns arrived. Bad years late in retirement are tolerable; bad years early can be catastrophic. Plan defensively for the early years; the late years usually take care of themselves.