Spending that grows with inflation, a portfolio that earns a steady yield, and a horizon of \(n\) years: under these assumptions the pot that is spent down to exactly zero has a closed formula. This note derives it, then lets you run your own numbers — including the year you can stop working.
The question
FIRE (Financial Independence, Retire Early) reduces to a single number: how much is enough? The popular 25× rule prices a near-perpetual portfolio. Here we answer a sharper version of the question: enough to fund exactly \(n\) more years of inflation-growing spending — no more, no less.
The model
Write \(S\) for this year's spending, \(i\) for inflation, \(r\) for the yield on investments, and \(R_i = 1+i\), \(R_y = 1+r\). Assume:
- a full year of spending is withdrawn at the start of each year; whatever remains stays invested and earns \(r\) over that year;
- spending inflates at \(i\): year \(t\) costs \(S\,R_i^{\,t-1}\) (year 1 is at today's prices);
- the yield \(r\) is the same every year;
- the money must last exactly \(n\) years — the pot reaches zero at the end of year \(n\), leaving nothing behind;
- while you are still working, spending is zero in the model: the yearly saving figure of §5 is what is left after all spending, so spending is already accounted for;
- savings are deposited at the end of each year, and the deposit grows with inflation (income is assumed to keep pace with prices).
From recurrence to formula
Start with \(P\); withdraw, let the rest grow, and repeat until nothing is left after \(n\) withdrawals:
$$\Bigl(\bigl((P-S)\,R_y - S R_i\bigr)R_y - S R_i^{2}\Bigr)R_y - \cdots \;=\; 0. \tag{1}$$Expanding the brackets:
$$P\,R_y^{\,n} \;-\; S\bigl(R_y^{\,n} + R_i R_y^{\,n-1} + R_i^{2} R_y^{\,n-2} + \cdots + R_i^{\,n-1} R_y\bigr) \;=\; 0. \tag{2}$$Divide through by \(R_y^{\,n}\): the sum turns geometric with ratio \(x = R_i / R_y\), and
If \(r = i\) then \(x = 1\) and (3) collapses to \(P = S\,n\): the yield only treads water against inflation, so you need the full \(n\) years of spending up front.
How much do you need today?
Table 1 — year-by-year drawdown for the parameters above.
| Year | Spending | After withdrawal | Interest | Value at year end |
|---|
When can you retire?
Now the saving side. Suppose you put aside \(D\) per year — already net of all spending, per assumption (e) — growing with inflation, deposited at year end, compounding at \(r\). Meanwhile the amount you would need in order to retire falls as the years still to be funded shrink: with \(n-t\) years left, plug \(n-t\) into (3):
$$\mathrm{Need}_t \;=\; S\cdot\frac{1-x^{\,n-t}}{1-x}. \tag{4}$$The first year your savings overtake \(\mathrm{Need}_t\) is your FIRE year — the highlighted row of Table 2. (\(\mathrm{Need}_t\) is stated in today's spending power; see remark 3.)
Table 2 — savings vs. the amount needed to retire (first sufficient year highlighted; rows stop a few years past the crossover).
| Year | Saved this year | Total saved (year end) | Needed to retire | Surplus |
|---|
Remarks
- \(r\) is assumed constant. Real portfolios deliver returns in a random order, and bad early years hurt a retiree disproportionately (sequence-of-returns risk) — prefer a conservative \(r\).
- \(i\) and \(r\) are nominal and before tax and fees; use your own net figures.
- \(\mathrm{Need}_t\) in Table 2 keeps spending at today's level. To state it in year-\(t\) dollars instead, multiply by \(R_i^{\,t}\) — the crossover then arrives a few years later.
- \(n\) is yours to choose: 70 suits a 30-year-old planning to age 100. Try 60 or 50 and watch \(P\) fall.
Model numbers, not financial advice.