Asset Allocation for Early Retirement Has an Equity Floor

Post the question in any retirement forum, how conservative can a FIRE portfolio get and still beat inflation, and the replies sort into two piles. One pile offers reassurance: a 60/40 mix is fine, it has always worked. The other delivers a sermon: stocks for the long run, volatility is the tuition. Neither pile supplies a threshold, a computable line where conservative stops working. The line exists. At a 4% withdrawal rate on historical US real returns it sits near 40 to 50% stocks, and it comes from arithmetic rather than temperament.

The mechanism fits on an envelope. A portfolio holds its real purchasing power only while its expected real return exceeds its inflation adjusted withdrawal rate. Every bond added to the mix dilutes expected return, and at some equity share the mix's return sinks to the withdrawal rate. Above that share, de-risking mostly trades upside for calmer early years. Below it, you swap a risk that might happen for a loss that almost certainly will, and a 50 year retirement gives that loss half a century to compound.

So this piece does what the threads don't: derive the floor from your own withdrawal rate, check it against the historical safe withdrawal rate research, price the perfectly safe alternative with real Treasury yields, and hand over a decision rule. Asset allocation for early retirement is usually argued from nerve. It divides cleanly instead.

The Real Question Behind Asset Allocation for Early Retirement

The question as usually asked, what is the most conservative stock bond mix that still beats inflation, contains a hidden error. The bar is higher than beating inflation: the portfolio must beat inflation plus its own withdrawal. Long term US government bonds compounded near 2% real over the past century, so an all bond portfolio did outpace CPI on average. It could not outpace CPI plus a 4% draw. On average it fell short of that combined hurdle by roughly 2% a year, which is a portfolio that beats inflation on paper while going broke in real terms.

The second hidden assumption is that safety rises steadily as the stock share falls. Safety in a withdrawal portfolio is non-monotonic. Cutting equities from 80% to 60% genuinely trades some upside for steadier early returns. Cutting equities below the floor trades market risk for scheduled inflation loss. Two questions follow, and the rest of this article answers both: where does the historical record say withdrawal sustainability turns down, and what does a genuinely risk free portfolio actually pay?

The Arithmetic That Sets the Equity Floor

Start by working in real terms. Subtract inflation from every return before comparing anything, because your withdrawal escalates with CPI and the only return that matters is the one left after inflation. On US data since 1926, large company stocks compounded at roughly 6.5 to 7% real and long term government bonds near 2% real, per Ibbotson's SBBI data. A century that includes a depression, a world war, and stagflation produced those averages, so treat them as a base case, not a promise.

A mix's expected real return is the weighted average of its parts. Take the 30/70 portfolio conservative savers keep asking about: 0.30 × 6.5% + 0.70 × 2% = 3.35% real. Against a 4% withdrawal, that mix falls behind its own payout by about 0.65% a year in expectation. With returns arriving perfectly evenly, a 3.35% real return paying 4% real depletes the portfolio around year 55. Volatility does not excuse the shortfall; it moves the depletion date around, typically earlier, because fixed withdrawals interact badly with early losses.

The floor is the equity share where the weighted average exactly equals the withdrawal rate. At historical inputs, 4% clears at about 44% stocks: 40% if you use 7% real equities, the high 40s if you assume intermediate bonds near 1.5% real. Drop the equity assumption to 5% real, a defensible forward view, and the 4% floor rises to 67%. One comparison carries all of this, real return versus withdrawal rate. When the first number is smaller, real erosion is the base case, not a tail event.

What the Historical Record Shows by Allocation

Bengen's original 1994 study rolled actual market history through 30 year retirements across a range of stock allocations and asked what initial withdrawal rate survived the worst starting dates. Two results matter here. The highest sustainable rates occurred at allocations between 50% and 75% stocks. And as the equity share fell below roughly 40 to 50%, sustainable rates declined, with an all bond portfolio sustaining materially less than 4% in the same study's tables.

Tabulations of that same rolling history, later popularized by Michael Kitces, sharpen the curve: the safe withdrawal rate by asset allocation peaks near a 60/40 mix and falls steadily as equities drop toward zero. Notice how well the arithmetic and the history reconcile. The 40 to 50% region is where the empirical curve sags, and it is also where the expected return of the mix falls to a 4% hurdle on historical inputs. The mechanism and the data point at the same line.

The historical peak sits above the floor because volatility is expensive. Expected return is what you would collect if returns arrived evenly, and they arrive in decade-long lumps instead, so durable allocations need margin above the arithmetic minimum.

That also settles the recurring question of whether bonds beat inflation in retirement. As an asset class, long bonds historically did, by about 2% a year. As a 4% withdrawal vehicle, they fell two points short. A stock bond mix meant to beat inflation over 40 years of drawdowns generally needs an equity majority.

Compute Your Equity Floor From Your Withdrawal Rate

This part takes ninety seconds in a spreadsheet. Everything downstream is just choosing inputs honestly.

The formula

equity floor = (w - b) / (s - b)

w = your real withdrawal rate
s = assumed real return on stocks
b = assumed real return on bonds

Set the stock share so that (stock share × s) + (bond share × b) equals the withdrawal rate, then solve. The answer is the minimum equity allocation at which the mix's expected real return still covers the payout.

Floors at five withdrawal rates

Real withdrawal rate	Floor at 6.5% stocks, 2% bonds	Floor at 5% stocks, 2% bonds
3.00%	22%	33%
3.25%	28%	42%
3.50%	33%	50%
3.75%	39%	58%
4.00%	44%	67%

Read the 4% row as the minimum stock allocation for the 4% rule: somewhere between 44% and 67%, depending entirely on which return assumptions you will defend. Small input changes matter. Use 7% real equities and the historical floor drops to 40%; use 1.5% real bonds and it climbs into the high 40s. In asset allocation for early retirement, the withdrawal rate and the return assumptions do all the work. The formula itself is just division.

Where the inputs come from, and two cautions

Don't borrow an assumption you can't defend. Damodaran's annual dataset gives stocks, bonds, and bills back to 1928, so you can recompute real returns over any window you consider relevant. Two cautions apply to the number itself. First, the floor is a necessary condition, not a sufficient one; it marks where expectation stops covering spending and says nothing about variance or ordering. Second, the floor moves with yields. At 1% real bonds every row rises; at 2.5% they ease.

Treat the floor as a basement, not a landing. History's durable allocations sit above it, because volatility spends part of the expected return. The floor tells you where de-risking stops being free, not where you should stop.

Why 40 to 60 Year Horizons Raise the Stakes

A longer horizon moves the floor itself, not just the odds. Every decade added to a retirement lowers the withdrawal rate the historical record will support, because a sub-floor shortfall compounds for more years. The floor formula takes that rate as input, so the floor sinks with it. Plug 3.25 to 3.5%, the rates long-horizon research certifies, into (w - b) / (s - b): historical inputs give 28 to 33% stocks, conservative inputs 42 to 50%, the same cells in the table above. The floor and the safe rate fall together, which is why a 50 year retirement remains survivable at all.

The Early Retirement Now series puts the safe withdrawal rate for a 50 year retirement at 3.25 to 3.5%, achieved mostly at equity shares of roughly 75% or higher. That is not a rival claim to the arithmetic but the sequence buffer stacked on the lower floor: the floor marks where a 3.25% payout is covered in expectation, and surviving six decades of bad sequencing demands margin above it.

The inputs deserve a haircut too. The US record is a survivor's record; the Global Investment Returns Yearbook tracks more than a century across dozens of markets, and global real averages run below the US figures. The rule that falls out is the opposite of intuition: at the same withdrawal rate, a 60 year plan needs more equity than a 30 year plan. That is also why the 30 year success rates of the Trinity study (Cooley, Hubbard, and Walz) cannot be quoted as evidence for a 60 year drawdown.

The TIPS Benchmark for a Perfectly Safe Portfolio

If your goal is to eliminate market risk entirely, the instrument exists. Treasury Inflation Protected Securities pay principal and coupons adjusted for CPI, and a ladder matched to your horizon delivers income escalating with the same index behind BLS CPI data. A TIPS ladder is the cleanest benchmark in finance, because it is where inflation risk in a retirement portfolio gets an explicit price.

A TIPS ladder's sustainable withdrawal rate is pure annuity arithmetic. A ladder paying level real income for T years at a locked real yield r supports:

sustainable real income = r / (1 - (1 + r)^-T)

r = real yield of the ladder, T = years of income needed

Real yield	40 year income	50 year income	60 year income
2.0%	3.66%	3.18%	2.88%
2.5%	3.98%	3.53%	3.24%

Long TIPS yields have traded around 2 to 2.5% real in recent years, so the table reads directly: a 4% inflation adjusted withdrawal exhausts a pure TIPS ladder in roughly 35 to 40 years, and funding a full 60 years requires withdrawing about 3% or less. So the TIPS ladder versus stocks for early retirement comes down to a price: everything above roughly 3% real for life has to be bought with stock risk.

One practical drag deserves a mention: in a taxable account, TIPS inflation adjustments are taxed as income as they accrue, which quietly lowers the real spendable yield the whole ladder depends on.

Sequence Risk Versus Slow Inflation Failure

The two ways a drawdown plan dies look similar in a spreadsheet and are opposites in character.

Sequence risk is probabilistic and front loaded

Sequence of returns risk is a coin flipped in your first decade. Fixed withdrawals from a falling market lock in losses that later recovery cannot fully repair, and Pfau's sequence research on the historical record concluded that retirement outcomes are dominated by returns over roughly the first decade of the drawdown. That is how two cohorts with identical average returns produce one survivor and one failure. The risk is probabilistic: most historical cohorts never drew on it, and a retiree ten years in with the portfolio intact has largely aged out of the danger zone.

Sub floor erosion is slow and near certain

Erosion below the floor is a bill that arrives on schedule. If the mix's expected real return sits under the withdrawal rate, the plan fails unless realized returns beat your own assumptions, and every extra year of horizon adds another year of shortfall. The loss compounds quietly and gets proportionally worse the longer the retirement, which is why 60 year outcomes concentrate at high equity shares while 30 year outcomes tolerate almost anything above the floor. At a 4% draw across 50 years, the 30/70 mix's failure is the near certain, scheduled one; the 70/30 mix's risk is the probabilistic, front loaded one.

A Decision Rule for Your Stock Bond Mix

The stock bond mix for retirement that fits you is an output of a procedure, not a preference. Five steps:

1. Fix the withdrawal rate first. It is the most powerful input you control, and everything else derives from it. For a 50 to 60 year horizon, anchor planning near 3.25 to 3.5% rather than the 30 year 4% figure.
2. Set real return assumptions, then haircut them. Historical inputs (6.5 to 7% real stocks, about 2% bonds) and conservative inputs (5% and 2%) bracket the plausible range; the global record argues for the lower end.
3. Compute the floor: (w - b) / (s - b). This is the minimum stock share at which expected return still covers spending.
4. Add a sequence buffer. The floor is where expectation covers the payout; history's durable allocations ran 50 to 75% stocks for 30 years and roughly 75% or more at 50 to 60. The buffer is the distance between arithmetic and history.
5. If floor plus buffer exceeds your risk tolerance, change the withdrawal. The reliable levers are a lower withdrawal rate, spending guardrails, or a TIPS layer funding core spending at about 3% real. Pushing the equity share below the floor is not on the list, because it substitutes a scheduled loss for a market risk.

The one respectable counterpoint is the research on the rising equity glide path, which found that retiring with a below target equity share and climbing toward it through retirement reduced worst case outcomes in many historical simulations. Read what that strategy actually does: it starts near these floors and rises, treating low equity as a temporary position against sequence risk rather than a permanent answer to inflation. Holding 30/70 for 50 years is a different thing entirely.

The forum question about how conservative a portfolio can get has an exact answer: down to your floor, and your floor is set by your withdrawal rate, not by your nerves. Below it, every additional bond converts market uncertainty into a near-certain loss with a computable arrival date. Recompute when yields or assumptions move. The number will drift. The rule holds: when safety feels too expensive at the floor, the honest move is to spend less, not to own fewer stocks.