Why should we invest our money?

One reason — if you can see the world a little more broadly, can calculate risk, and can plan long-term, you can see that this rat race actually has an exit.

Let me do a very dry calculation. Kids, big houses, luxury cars are the biggest liabilities — I’m leaving those out of the core calculation, with a brief with-kids variation at the end.

Annual cost of living: Seattle

How much do you need per year to live comfortably? Take Seattle, where I am, as the example.

Category $/month
Rent (1-bedroom, non-luxury) 2,000
Food 700
Car (insurance, gas, maintenance, amortized depreciation) 500
Utilities (electricity, gas, water) 150
Communications (internet + mobile) 150
Healthcare (employer plan: premium share + out-of-pocket) 150
Personal / discretionary 500
Travel + misc (annualized) 500
Effective monthly 4,650

Annual: about $56k/year during working years. The $150/month healthcare line is what you pay on a typical employer plan — premium share + average out-of-pocket. An employer subsidy reduces the cost a lot, but it doesn’t make it zero. In retirement you lose the employer subsidy entirely and need to buy coverage on the ACA marketplace — closer to $500/month for a healthy 30-something — so retirement expenses come to about $60k/year.

Yes, you need a car in Seattle. Public transit covers the urban core, but enough of daily life sits outside it that car-free is impractical for most people.

So to escape the rat race here, you need investment returns alone to cover ~$60k/year. Let’s call the point where investment returns alone cover your annual expenses the rat-race escape zone. It is commonly called early retirement, or FIRE.

The S&P 500 as the investment baseline

Buying individual stocks is too risky — diversify across a broad-market index ETF. The popular index is the S&P 500. It goes back to 1926; has lived through the Great Depression, the 1970s stagflation, the dot-com crash, 2008, COVID, and every recession in between. With dividends reinvested, the annualized nominal return from 1926 to 2024 averages ~10.3%, and the post-1957 era (since the modern 500-stock form) averages ~10.5%. Inflation has averaged ~3% over the same period, so the real return — what actually matters for purchasing power — is closer to ~7%. A century of history validates this.

The past 50 years specifically (~1976 onward) have been more favorable still — about 11.9% nominal and 8.2% real — partly because that window happens to skip the stagnant, high-inflation late-1960s and early-1970s. For planning purposes I’ll stick with the full-century ~7% real baseline. It’s the more conservative anchor — including the worst stretches as well as the best — and that’s the right side to err on when the math is forecasting decades into the future.

Why the safe withdrawal rate is much lower than the return

You might be tempted to plug ~10% — or ~7% real — directly into the math as your withdrawal rate. That’s wrong.

The average return is what the index does over time. The maximum amount you can withdraw each year in retirement is much lower. Two reasons:

  1. The horizon problem. A 30-year retirement is shorter than the “long enough to mean-revert” window. The long-term average assumes you can wait out any downturn. A retiree drawing from the portfolio cannot.

  2. Sequence-of-returns risk. Imagine two retirees with identical $1M portfolios, both drawing the same inflation-adjusted amount per year, both experiencing the same set of annual returns over 30 years — but in reverse order. The arithmetic average is identical. The retiree who hits the bad decade first runs out of money. The other dies wealthy. Returns are not commutative once you start making withdrawals.

This is the reason the safe withdrawal rate is so much lower than the average return. The well-known answer is the 4% rule — Bengen (1994) and the Trinity Study (1998), derived by backtesting U.S. history to find the maximum constant inflation-adjusted withdrawal that survived every 30-year window, including the worst sequences. Modern revisits push this even lower; Wade Pfau and others argue 3.0–3.5% is more defensible given today’s elevated equity valuations and structurally lower bond yields.

For this game plan I’ll use 3.5%. It’s a personal judgment.

Seattle: how long to the rat-race escape zone

So the rat-race escape-zone target for a Seattle retirement:

rat-race escape zone=$60,0000.035≈$1.7M\text{rat-race escape zone} = \frac{\$60{,}000}{0.035} \approx \$1.7\text{M}

We use the compound-interest formula with monthly contributions to calculate how long it takes:

target=P⋅(1+r12)12t  +  income⋅(1−tax)−expenses12⋅(1+r12)12t−1r/12\text{target} = P \cdot \left(1 + \tfrac{r}{12}\right)^{12t} \;+\; \frac{\text{income} \cdot (1 - \text{tax}) - \text{expenses}}{12} \cdot \frac{\left(1 + \tfrac{r}{12}\right)^{12t} - 1}{r/12}

Where PP is starting capital, rr is the real return (not nominal — I’m using real, so the target stays in today’s dollars and inflation drops out), tt is years.

Quick aside on how the formula works — it’s two compound-growth streams added together:

  • Your starting capital term, P⋅(1+r/12)12tP \cdot (1+r/12)^{12t}, is PP compounding monthly at rate r/12r/12 for 12t12t months. Standard “money grows at interest” math.

  • Your contributions term sums the future value of each monthly contribution. Each month you add x/12x/12 (annual saving rate divided by 12). Each chunk sits in the market for a different number of months before time tt — the first month’s chunk has 12t−112t-1 months to grow, the last has zero. So the total is

    x12⋅[(1+r/12)12t−1+(1+r/12)12t−2+⋯+(1+r/12)0]\frac{x}{12} \cdot \left[ (1 + r/12)^{12t-1} + (1 + r/12)^{12t-2} + \cdots + (1 + r/12)^0 \right]

    That bracketed geometric series has the closed form (1+r/12)12t−1r/12\frac{(1+r/12)^{12t} - 1}{r/12} — the fraction in the formula.

Both streams compound at the same monthly rate; you add them to get your portfolio value at time tt.

Assume you start working at age 22, straight out of college, as a software engineer in Seattle. Most fresh grads don’t have meaningful starting capital, so we’ll set P=0P = 0 — the formula collapses to just the contributions term.

  • Career-average compensation income: $200k (representative of a Seattle engineer averaging across their first 10–15 years, growing from a new-grad $130–150k to senior $250k+)
  • Effective tax: 20% (federal only — WA has no state income tax)
  • Working-years expenses: $56k
  • Real return r: 7% (matches the ~7% historical S&P 500 real return for a long-horizon equity-heavy accumulation phase. We’ll drop to the lower 3.5% withdrawal rate once retired.)
  • Target: $1.7M

Plugging in, the equation we need to solve is:

200,000⋅0.80−56,00012⋅(1+0.0712)12t−10.07/12=1,700,000\frac{200{,}000 \cdot 0.80 - 56{,}000}{12} \cdot \frac{\left(1 + \tfrac{0.07}{12}\right)^{12t} - 1}{0.07/12} = 1{,}700{,}000

Solve on Wolfram Alpha →

Result: t ≈ 11 years. Rat-race escape zone at age 33.

Visualizing both trajectories — same $104k/year savings, one invested at 7% real return, one just sitting in cash. The compounding effect bends the invested line steeper over time; the cash line stays linear (all values in today’s dollars):

%%{init: {"themeVariables": {"xyChart": {"plotColorPalette": "#c62828,#000000"}}}}%%
xychart-beta
    title "Portfolio growth at $104k/year savings (inflation-adjusted)"
    x-axis "Years" 0 --> 12
    y-axis "Portfolio ($M)" 0 --> 2
    line [0, 0.11, 0.22, 0.35, 0.48, 0.62, 0.77, 0.94, 1.11, 1.30, 1.50, 1.72, 1.95]
    line [0, 0.10, 0.21, 0.31, 0.42, 0.52, 0.62, 0.73, 0.83, 0.94, 1.04, 1.14, 1.25]

Red = invested at 7% real return. Black = not invested — cash held in something that at least matches inflation (a high-yield savings account, money-market fund, or short-term Treasuries). The invested curve hits the cross-border target ($800k) at year ~6 and the Seattle target ($1.7M) at year ~11.

The black line assumes the minimum responsible default — your cash keeps up with inflation. If it sits in a checking account at ~0% while inflation runs 2–3%, real value actually erodes year over year and the line bends downward. Doing nothing isn’t neutral; it costs purchasing power.

A quiet bonus: withdrawal-phase tax is essentially zero

The $60k target above is what you need to spend — it implicitly assumes the withdrawal doesn’t have to be grossed up for tax. For an early retiree pulling from a long-held S&P 500 index in a taxable brokerage account, that’s correct, and it’s because of how the U.S. taxes long-term capital gains.

LTCG (positions held >1 year) have their own brackets. The bottom one pays 0% federal tax. For tax year 2026, the 0% LTCG bracket covers taxable income up to $49,450 single / $98,900 MFJ — and “taxable” means after the standard deduction ($16,100 / $32,200). So you can realize roughly ~$65k single / ~$131k MFJ of gross long-term capital gains at $0 federal tax. The thresholds are inflation-adjusted annually.

A single Seattle retiree pulling $60k/year sits below the ceiling. And only the gain portion of each sale counts — if your $60k came from shares with a $40k cost basis, only $20k hits the LTCG total. Federal: $0. WA state: $0. Effective withdrawal-phase tax rate: 0%. The plan doesn’t just reach the rat-race escape zone — it stays tax-free below it indefinitely.

An alternative — retire outside the US

The Seattle math assumes you stay in Seattle through retirement. But the cost-of-living gap between U.S. cities and many comfortable places to live elsewhere is large enough that the retirement location is, on its own, one of the biggest levers in this equation.

Take Tokyo as the example, since it’s the one I have direct context on. A comfortable single-person budget in central Tokyo today:

Category ¥/month
Rent (1K/1LDK, central-ish) 120,000
Utilities (electricity, gas, water) 15,000
Communications (internet + mobile) 15,000
Transit (no car) 15,000
Healthcare (national plan + out-of-pocket) 30,000
Food 60,000
Personal / discretionary 50,000
Travel + miscellaneous (annualized) 50,000
Effective monthly 355,000

Annual: ~¥4.2M, which at ¥150/$ is about $28k/year USD — well under half the Seattle equivalent. Two big reasons for the gap: Tokyo doesn’t require a car (the train network actually works), and Japan’s national health insurance is meaningfully cheaper for a retiree (~¥30k/month, scaled to retirement-level income) than the $500/month ACA-marketplace line in Seattle.

Target in USD at the same 3.5% withdrawal rate:

rat-race escape zone=$28,0000.035≈$800k\text{rat-race escape zone} = \frac{\$28{,}000}{0.035} \approx \$800\text{k}

Same Seattle-engineer income and savings rate, much smaller target:

200,000⋅0.80−56,00012⋅(1+0.0712)12t−10.07/12=800,000\frac{200{,}000 \cdot 0.80 - 56{,}000}{12} \cdot \frac{\left(1 + \tfrac{0.07}{12}\right)^{12t} - 1}{0.07/12} = 800{,}000

Solve on Wolfram Alpha →

Result: t ≈ 6 years. Rat-race escape zone at age 28.

Earning at U.S. tech salaries and spending at Japanese cost of living cuts the timeline almost in half. The principle generalizes — you can run the same math with Lisbon, Chiang Mai, Mexico City, or any number of other lower-cost destinations. The specifics change, the shape doesn’t.

One caveat worth flagging on the financial side — currency risk. ¥150/$ is the rate today; the yen is historically weak right now. If it strengthens to ¥100/$ (the average from 2010–2020), your $28k/year budget becomes $42k/year and the target balloons past $1.2M. Either hedge or carry a larger buffer than the headline number suggests. Moving logistics and immigration are out of scope for this post.

Tokyo native: same target, much longer path

For engineers reading this in Tokyo who plan to stay — the retirement target is the same physical amount as the cross-border path, just expressed in yen. At ¥150/$, ¥120M is $800k — same number, different currency. What changes is the savings rate: Japanese tech salaries vs U.S. tech salaries. ¥9M career-average for a mid-senior engineer, 22% effective tax, ¥7.02M take-home, ¥4.2M expenses, so ¥2.82M/year going into the portfolio. Starting from zero at age 22:

9,000,000⋅0.78−4,200,00012⋅(1+0.0712)12t−10.07/12=120,000,000\frac{9{,}000{,}000 \cdot 0.78 - 4{,}200{,}000}{12} \cdot \frac{\left(1 + \tfrac{0.07}{12}\right)^{12t} - 1}{0.07/12} = 120{,}000{,}000

Solve on Wolfram Alpha →

Result: t ≈ 20 years. Rat-race escape zone at age 42.

The all-Japan path is more than three times longer than the Seattle-to-Tokyo path. Software engineer compensation in Japan relative to local cost of living simply isn’t where U.S. tech is.

And that assumes you actually invest, which a lot of Japanese people don’t. Households here hold roughly half of net financial wealth in cash and deposits (vs. about 15% in U.S. households). Two historical reasons: in the 1970s, Japan’s post office paid roughly 7% on plain savings — for an entire generation, just save in the post office was a perfectly rational wealth-building strategy. Then the late-1980s asset bubble peaked: the Nikkei 225 hit 38,915 on December 29, 1989, and didn’t recover that level until February 2024 — more than 34 years later. A generation that did invest watched their portfolios stay underwater for three decades. The lesson stuck: investing is dangerous, saving is safe. That lesson is now backwards (rates near zero, inflation back to 2–3%), but the cultural memory persists.

The math today says save aggressively and invest, not save or invest.

What this math doesn’t cover

Two foundational assumptions the math quietly depends on: no excessive debt and staying healthy. The income side assumes you can keep earning; the savings side assumes a positive savings rate. Consumer debt — credit cards, car loans, student loans dragged into your 30s — destroys the savings rate. Burnout, chronic illness, or a serious mental-health stretch destroys the income side. Live within your means, never spend more than you earn, and take care of yourself mentally and physically. The math is only as good as those two foundations.

Life has other ups and downs too: medical events, divorce, lifestyle inflation, a market crash early in retirement. Any of these can stretch the timeline significantly.

The biggest single asterisk is kids. Rule of thumb: a family of four roughly doubles the expenses above — Seattle family ~$112k/year working, ~$120k/year in retirement (¥8.4M/year in Tokyo) — which roughly doubles the rat-race escape-zone target too. Seattle becomes about $3.4M instead of $1.7M. The timeline doesn’t double automatically, though. Two Seattle engineers earning $200k each contribute ~$208k/year combined, and hit the new $3.4M target in ~11 years — essentially the single-person timeline. The math is approximately scale-invariant when both partners earn similarly. The timeline blows up when the household drops to a single income: one $200k earner with $112k family expenses contributes only ~$48k/year, and the same $3.4M target now takes ~26 years. Same target, very different timeline.

We can’t predict life’s ups and downs, but knowing what’s possible — rat-race escape zone in your late 20s with the cross-border path, or early 30s in Seattle alone — is itself hopeful. The math is the easy part. The hard part is the decades of discipline in between: staying invested through downturns, not lifestyle-inflating, not raiding the portfolio for one-off “I deserve it” purchases. Most people who don’t reach the rat-race escape zone don’t fail at math. They fail at the discipline.

If you’re earlier in your career: start now, automate it, and don’t touch it during downturns. The rest is bookkeeping.

After escape, you can keep working because you actually want to, volunteer, play FPS all day, live on a beach in Okinawa — your call. Just not by obligation.

Read more like this