Safe withdrawal rates explained

What is a Safe Withdrawal Rate (SWR)?

A Safe Withdrawal Rate is the percentage of your starting portfolio you can withdraw in year 1, then increase that amount each year with inflation, with a high likelihood of not running out over your chosen horizon.

The canonical example: 4% SWR on ₹3 crore = ₹12 lakh year 1, then ₹12.6 lakh if inflation is 5% in year 2, etc. If this survived the worst historical periods in testing, it’s called “safe” for that dataset.

How SWR is derived

Researchers (Bengen, Trinity study) tested:

1. Starting portfolio split between stocks and bonds (e.g. 60/40).
2. Fixed inflation-adjusted withdrawals each year.
3. Historical return sequences (US data from 1926 onward).
4. Find the highest rate that worked for all starting years for a given horizon.

The result: for 30-year horizons, 4% was “safe” historically. For shorter horizons, higher rates worked. For longer horizons (early retirement), lower rates were needed.

SWR by horizon and conservatism

Horizon	Conservative (high confidence)	Moderate	More aggressive
15–20 years	4.5–5.0%	5.0–5.5%	5.5%+ (less safe)
25 years	3.8–4.0%	4.0–4.2%	4.3% (tight)
30 years	3.5–3.75%	3.75–4.0%	4.0% (benchmark)
35–40 years	3.25–3.5%	3.5–3.6%	3.7%+ (risky)

These ranges reflect modern research (Morningstar, Pfau, Kitces) accounting for lower real yields.

SWR vs Success Rate

There’s an important distinction:

• SWR (historical “safe”): “Would this have worked in the worst 30-year period in US history?”
• Monte Carlo success rate: “Given our return/inflation distribution, what % of 10,000 simulated paths survive?”

A 4% rate might be “100% safe” historically, but only 85–92% successful in Monte Carlo depending on assumptions. Neither is “guaranteed”. Monte Carlo is often more informative because it tests many more paths than just the ~100 years of history we have.

Use our Monte Carlo simulator to see probabilities for your specific numbers.

Why the 4% rule isn’t universal

• Low bond yields. Starting yields are a strong predictor of future real returns. Today’s yields are lower than the 1980s–90s, dragging expected returns.
• Longer retirements. Early retirees need 40+ years, not 30. The safe rate drops.
• Global markets differ. US outperformed many markets historically. A globally diversified portfolio still needs care.
• Fees eat returns. 1% fees compounds to a big drag over 30–40 years.
• Inflation regime matters. The 1970s stagflation was the real stress test. If inflation is higher and sustained, 4% is tighter.
• Valuations. High CAPE ratios have historically been associated with lower forward returns.

Beyond the rigid 4%: dynamic approaches

Static inflation-adjusted withdrawals are simple but not optimal. Better approaches adapt:

Approach	How it works	Pros	Cons
Percentage-of-portfolio	Withdraw X% of current value each year (e.g. 4% of remaining).	Never mathematically runs out; naturally cuts in down years.	Spending can be very volatile year-to-year.
Guyton–Klinger guardrails	Inflation-adjust, but cut if portfolio down > 10–20%; freeze/cut if over-withdrawn relative to rules.	Higher lifetime spending, similar safety; adapts to reality.	More rules to follow; feels counterintuitive in a crash.
Floor & ceiling	Inflation-adjust, but never increase > ceiling (e.g. +10% over inflation), never cut below floor.	Smooths spending; prevents big cuts.	Slightly more complex.
VPW (Variable Percentage Withdrawal)	Uses life expectancy tables to recalc percentage each year.	Adjusts for remaining horizon; never guaranteed to last but adapts.	Can feel technical; sensitive to life expectancy assumptions.
Bucket strategy	Spend from cash bucket in down years; refill from growth in good years.	Very intuitive; reduces sequence risk psychologically.	Needs discipline to refill properly.

Research consistently shows dynamic rules allow higher average spending with equal or better safety than rigid 4%.

A practical way to choose your SWR

1. Pick your horizon. If retiring at 65, 30 years is fine. If FIRE at 45, use 35–40 years.
2. Assess flexibility. Can you cut discretionary by 10–20% in a bad year? If yes, you can be closer to 4%. If no (fixed costs only), be more conservative.
3. Start conservative. Try 3.5–3.75% first. See the required corpus. If it feels daunting, test 4% but also test a dynamic plan.
4. Stress-test. Run Monte Carlo (Monte Carlo simulator) and test a bad-early sequence (Sequence of returns).
5. Plan for a buffer. 1–2 years cash + an emergency fund makes even 4% more resilient.
6. Review periodically. Your SWR isn’t locked in. At 70, with a smaller horizon, you could safely increase withdrawals.

Common misconceptions

• “4% is guaranteed.” No. It survived US history, but future could be worse. Also, behavioural mistakes break any rule.
• “Higher equity % always means higher SWR.” Not quite. Very high equity can increase worst-case failures in some sequences. 40–70% equity is the usual sweet spot.
• “Once set, never change.” Rigid is brittle. The most successful retirees adjust spending based on reality.
• “SWR is the same globally.” Different markets, currencies and inflation regimes mean you should be a bit more conservative if not globally diversified.
• “25× is enough for everyone.” Only if your spending estimate, horizon, flexibility and assumptions match.

Worked example

Case A: Retire at 65, 30-year horizon, flexible spending, globally diversified 60/40. SWR ~3.75–4.0% is reasonable.

Case B: FIRE at 45, 40-year horizon, want very high confidence, somewhat inflexible. SWR ~3.25–3.5% is more prudent.

Case C: Semi-retire at 55, have a small pension covering essentials, flexible wants. SWR ~4.0% on the discretionary portion is often fine.

Bottom line

Treat 3.75% as your default unless you have strong reasons otherwise. If you have spending flexibility, a cash buffer, or other income, 4% can be workable. If your horizon is very long or you’re risk-averse, lean towards 3.5%. Most importantly: don’t ask “is this safe?” — ask “how likely is this to work, and what levers can I pull if it doesn’t?” Use probability tools, not just historical rules.