Safe Withdrawal Rates Explained

The 4% rule is the most cited number in retirement planning and one of the most misunderstood. It was never intended as universal advice. It emerged from specific research, under specific assumptions, for a specific retirement length. Most early retirees — and many traditional retirees — should not be using 4% as their withdrawal rate. This guide walks through the actual research, what it found, what it did not find, and how to determine a withdrawal rate that fits your situation.

What “safe withdrawal rate” actually means

A safe withdrawal rate (SWR) is the percentage of your portfolio you withdraw in year one of retirement, then adjust for inflation each subsequent year, without running out of money over a specified period. The “safe” part means it survived every historical period tested — including the Great Depression, 1970s stagflation, and the dot-com crash.

Three variables define any SWR: the retirement length (how many years the portfolio must last), the portfolio allocation (stocks versus bonds versus other assets), and the definition of “safe” (100% success rate? 95%? 90%?). Change any of these and the rate changes. This is why “the 4% rule” as a universal constant is misleading — it only applies to one specific combination of these variables.

The Bengen research (1994)

William Bengen published “Determining Withdrawal Rates Using Historical Data” in the October 1994 Journal of Financial Planning. The paper tested every possible 30-year retirement period from 1926 onward, using a 50/50 portfolio of S&P 500 stocks and intermediate-term US Treasury bonds. For each starting year, he calculated the maximum withdrawal rate that would not have depleted the portfolio.

His finding: 4% survived every 30-year period. The worst starting year was 1966, when high inflation and poor stock returns created a particularly hostile environment for retirees. A 4% withdrawal rate starting in 1966 would have barely survived — the portfolio was nearly depleted by year 30.

The title of the paper matters: “Determining Withdrawal Rates.” Bengen was finding the ceiling — the maximum rate that never failed. He was not recommending 4% as a target. He was saying it was the highest rate you could have used without ever running out over any 30-year historical period. The distinction between “maximum survivable rate” and “recommended rate” has been lost in decades of popular retelling.

The Trinity Study (1998)

Four years later, professors Philip Cooley, Carl Hubbard, and Daniel Walz at Trinity University extended Bengen’s work. Their paper, “Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable,” tested multiple withdrawal rates across multiple portfolio allocations and multiple time horizons.

Key findings from the Trinity Study for 30-year periods: At 4% withdrawal with 50/50 stocks/bonds, the success rate was 95%. At 75/25 stocks/bonds, success rose to 98%. At a 5% withdrawal rate with 50/50, success dropped to 76%. At 75/25, 5% succeeded 83% of the time. At 6%, most allocations succeeded less than 60% of the time.

The Trinity Study gave us success probabilities across a matrix of rates and allocations. But it still used 30-year periods and US-only data. Someone retiring at 40 needs 50+ years, not 30. And someone relying on this research is implicitly betting that the next 50 years of US market returns resemble the last 100.

Why the 4% rule is not as safe as it sounds

Calibrated to 30-year retirements. The original research tested 30-year periods because that was the standard retirement horizon for someone retiring at 65 and living to 95. Early retirees need 40, 50, or even 60 years. At a 4% withdrawal rate over 50 years, historical success drops to roughly 85–90% depending on allocation. At 60 years, lower still. Longer retirements need lower rates.

Assumes US-only returns. The United States had an exceptional 20th century. Two world wars that devastated competitors while leaving US industry intact. The rise of global reserve currency status. Unprecedented technological innovation. Other developed countries — Japan, UK, Germany, France — had meaningfully lower stock market returns. Wade Pfau’s research on international SWRs found that safe rates for many non-US countries would have been 3% or lower. Banking on US exceptionalism continuing indefinitely is a bet, not a guarantee.

Assumes fixed inflation adjustment. Both Bengen and Trinity assumed retirees withdraw a fixed percentage in year one, then adjust that dollar amount upward by CPI every year regardless of portfolio performance. Real retirees have flexibility. If the market drops 40%, most people voluntarily reduce spending. This flexibility is valuable and not captured in the original research.

Sequence of returns risk

This is the concept that makes safe withdrawal rates necessary in the first place, and it is the single most important risk in retirement portfolio management.

Two retirees can experience identical average returns over 30 years and have completely different outcomes. The difference is when the returns occur. Poor returns in early retirement years, combined with ongoing withdrawals, create a double hit: the portfolio drops in value AND you are removing money from it. By the time returns improve, the portfolio may be too depleted to recover.

Example: $1,000,000 portfolio, $40,000/year withdrawal (4%). Scenario A: market drops 30% in year one, then recovers steadily at 10% for 29 years. Scenario B: market rises steadily at 10% for 29 years, then drops 30% in year 30. Both have identical average returns. In Scenario A, the portfolio ends with roughly $2.1M. In Scenario B, it ends with roughly $7.8M. Same returns, different sequence, $5.7M difference.

This is what the SWR simulator models: not average outcomes, but the full range of possible sequences and how they affect your specific portfolio.

What modern research suggests

The 4% rule was a starting point. Thirty years of subsequent research has refined it considerably.

Kitces variable withdrawal / Guyton-Klinger guardrails. Michael Kitces and Jonathan Guyton developed approaches where withdrawal rates adjust based on portfolio performance. The basic idea: set spending floors and ceilings. If the portfolio grows beyond a threshold, increase spending. If it drops below another threshold, reduce spending. These guardrail approaches allow initial withdrawal rates of 5.0–5.5% while maintaining high success rates — at the cost of 10–25% budget variability in bad years.

Big ERN Safe Withdrawal Rate Series. Karsten Jeske (Big ERN) published a 50+ part series analyzing withdrawal rates specifically for early retirees. His conclusions for 50-year retirements: 3.25–3.50% is the safe floor for a 75/25 stock/bond portfolio with near-100% historical success. His work specifically addresses the longer horizons that Bengen and Trinity did not test.

Wade Pfau funded ratio approach. Pfau reframes the question: instead of “what percentage can I withdraw?,” ask “what is my funded ratio?” — the ratio of current portfolio to remaining lifetime spending needs. A funded ratio above 1.0 means you have enough. This approach is more intuitive and adapts naturally to changing circumstances.

What rate should you actually use?

There is no single right answer, but there are defensible ranges based on retirement length and risk tolerance.

3.0%: Very conservative. Appropriate for 50+ year retirements, no flexibility in spending, no other income sources. Historically 100% success across virtually all periods and allocations. Requires $33.33 per dollar of annual spending ($2M for $60K/year).

3.5%: Moderate-conservative. Appropriate for 40–50 year retirements with some spending flexibility. Historical success above 95%. Requires $28.57 per dollar ($1,714,000 for $60K/year).

4.0%: Standard Bengen/Trinity rate. Appropriate for 30-year retirements. 95–100% historical success depending on allocation. Requires $25 per dollar ($1.5M for $60K/year).

4.5%: Moderate risk. Appropriate for 25–30 year retirements with backup income (Social Security, pension) or willingness to reduce spending in bad years. Historical success roughly 85–90% for 30-year periods. Requires $22.22 per dollar ($1,333,000 for $60K/year).

5.0%+: High risk without variable spending strategies. Only appropriate with guardrail rules, significant other income, or willingness to return to work if markets underperform. Requires $20 per dollar ($1.2M for $60K/year). The math works with Guyton- Klinger guardrails but demands budget discipline in down years.

The variable withdrawal question

Fixed withdrawal rates — set a dollar amount in year one and adjust for inflation forever — are the easiest to model and the worst to live with. In reality, almost nobody does this. People spend more when markets are up, less when they are down. Formalizing this flexibility produces better outcomes.

Guyton-Klinger guardrails. Set an initial withdrawal rate (say 5%). If the current-year withdrawal rate rises above 6% of portfolio (because the portfolio has shrunk), cut spending by 10%. If it drops below 4% (because the portfolio has grown), increase spending by 10%. The guardrails keep spending within a corridor and dramatically improve portfolio survival. The tradeoff: in the worst 10% of scenarios, you spend 20–25% less than planned for several years.

Floor-and-ceiling approach. Set a spending floor (essential expenses: housing, food, healthcare, utilities — perhaps $40,000) and a ceiling (full lifestyle including travel, dining, hobbies — perhaps $70,000). Withdraw ceiling in good years, floor in bad years. The portfolio only needs to reliably cover the floor; the ceiling is aspirational.

The spending smile. Research by David Blanchett found that retiree spending is not constant — it follows a “smile” pattern. Spending is high in early retirement (travel, activity), declines through the middle years, then rises again in late retirement (healthcare). Accounting for this pattern means early withdrawal rates can be slightly higher because spending naturally declines in the middle years.

The withdrawal rate simulator lets you test both fixed and variable strategies against historical data. Try your planned rate and see how it performs across different starting years.

The multi-source income advantage

Most SWR research assumes the portfolio is the sole income source. For many retirees, it is not. Social Security, pensions, VA disability compensation, rental income, and international pensions all create income floors that the portfolio does not need to cover. This changes the withdrawal rate question fundamentally.

If your expenses are $80,000/year and Social Security covers $30,000, the portfolio only needs to produce $50,000. On a $1.5M portfolio, that is a 3.3% withdrawal rate — well within even the most conservative safe range. The guaranteed income floor also provides psychological resilience during drawdowns: even if the portfolio drops 40%, essential expenses are covered by sources that do not fluctuate with markets.

This is why calculating your actual withdrawal rate — net of other income — is more important than debating whether 3.5% or 4% is “right.” Many retirees with moderate portfolios and robust income floors are effectively withdrawing 2–3% from their investments, which is safe under virtually any historical scenario. The withdrawal rate simulator accounts for additional income sources in its projections.

Core research for deeper reading

Bengen, 1994: “Determining Withdrawal Rates Using Historical Data,” Journal of Financial Planning. The original 4% rule paper. Available through the Journal of Financial Planning archives.

Trinity Study, 1998: Cooley, Hubbard, and Walz, “Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable,” AAII Journal. Extended Bengen’s work across allocations and time periods.

Kitces, Nerd’s Eye View: kitces.com. Ongoing research on variable withdrawal strategies, guardrails, and dynamic retirement income planning. The most comprehensive practitioner-oriented resource on withdrawal rates.

Big ERN, Early Retirement Now SWR Series: earlyretirementnow.com. 50+ part deep-dive specifically addressing early retirement withdrawal rates for 40–60 year horizons. The most rigorous analysis of SWR for FIRE specifically.

Pfau, Retirement Researcher: retirementresearcher.com. Wade Pfau’s work on funded ratios, international SWR data, and the retirement income planning framework. Essential for understanding why US-centric research may overstate safety.

This article is for educational purposes. It is not financial advice. Historical returns do not guarantee future outcomes. Withdrawal rate research is based on US historical data and may not apply to other markets or future conditions. Consult a fee-only financial planner for decisions specific to your situation.