Why 1965 Beats 1929: A 40-Year-Old's Data-Driven Reality Check on Lean FIRE
- Age
- 40
- Net Worth
- $1,000,000
- Expenses
- $40,000/yr
- FIRE Status
- On Track
Summary
This 40-year-old developer analyzed historical market sequences to prove that a 3.5% withdrawal rate is the gold standard for long-term lean FIRE success.
FIRE Analysis
Strengths
- • Your data-driven approach to stress-testing with 1965-style stagflation scenarios provides an excellent foundation for your long-term planning.
Areas to Focus
- • Incorporate a more formal variable-spending guardrail strategy into your actual portfolio to handle potential sequence-of-returns risks.
Potential Risk
- ⚠️ The high failure rate of the 4% rule during long-term horizons (45+ years) poses a significant risk to your retirement stability.
The Original Post
The 4% rule comes from the Trinity Study, which used 30-year retirements. Lean FIRE retirements are often 45-50 years. Does the math hold up at that horizon, especially after a bad sequence-of-returns? I wanted a clearer answer than "well, probably, with margin," so I built a tool that runs historical Monte Carlo against every actual 1928–2025 sequence — not random Monte Carlo, the actual market history. Here's what I found. Scenario I tested: * Lean retiree, age 40, **$1M portfolio** (real $) * **$40K/yr real spend** (= 4.0% starting withdrawal rate) * 60/35/5 stocks/bonds/cash allocation * 45-year retirement horizon * No Social Security factored in (baseline) * Flat draw — no variable-spending guardrail (the naive 4% rule) **Baseline result: 83.3% success** across 54 historical 45-year windows. The biggest surprise: failures cluster at 1965, 1966, 1968, 1969 — NOT 1929. Stagflation-era starts are the real killer for long lean retirements, not the Great Depression. The 1966 starter ran out of money at year 27 (age 67). Long horizon + bad first-decade returns + sticky inflation = the unique hell of late-60s retirees. *Then I layered in real-world guardrails:* Add Social Security at FRA (67): success rate jumps to 92.6%. I used $16,762/yr real — the SSA bend-point estimate for a $50K career earner who worked 18 years (age 22-40) before retiring. That's a "tiny" SS check by mainstream standards, but it makes a meaningful difference because it kicks in right when the bad-sequence portfolios are at their lowest. Add a conditional variable-spending guardrail (if $15K of your $40K spend is discretionary, cut it 50% in years following a market drop — Guyton-Klinger style): success rate hits 98.1%. Only the 1966 start still fails, and even that delays from year 27 to year 41. Drop spend to $35K/yr (3.5% WR) with SS at 67: **100% historical success.** Takeaways for lean FIRE specifically: * The 4% rule on a 45-year horizon isn't a sure thing. \~17% failure rate at $40K/yr on $1M with no guardrails. Worth knowing. * The conventional "1929 is the worst" wisdom is wrong for 45-year windows. The 1965-1969 stagflation starts are MUCH worse because there's no time to recover. * Social Security alone (even a "tiny" $16K/yr) lifts success \~10 percentage points. Lean folks should factor it in honestly, not dismiss it. * Token haircuts ($3K/yr cut) barely move the needle. Meaningful cuts (cutting your discretionary 30-50% in down years) actually save you. * 3.5% WR + SS = bulletproof historically. The "less aggressive" SWR isn't paranoia; it's the price of a 45-year horizon. Made the tool free, no signup, runs entirely in your browser. No analytics, no telemetry — you can open DevTools to verify. **Mobile Website:** [**https://wealthtrajectory.vercel.app**](https://wealthtrajectory.vercel.app) Source available under PolyForm NC (free for personal use): [https://github.com/vsriram11/wealthtrajectory](https://github.com/vsriram11/wealthtrajectory) If you want to stress-test your own lean plan against the same 1928–2025 sequences, you can plug in your numbers and run it. Honest feedback welcome, particularly: * Is the analytical setup realistic for typical lean FIRE folks? * What scenarios should I add to the historical MC engine? * Anything else you wish to customize beyond per-asset CAGRs, and allocations? * How would you extend the math? Built it because I wanted to know my own sequence risk precisely. Sharing in case it's useful here.