Warren Buffett Portfolio: The 90/10 Allocation Explained

What is the Warren Buffett portfolio?

The phrase usually refers to something very specific, and it is not Berkshire Hathaway's stock holdings. In his 2013 letter to Berkshire shareholders, Warren Buffett described the instructions he left for the trustee who will manage the cash going to his wife: put 10% of it in short-term government bonds and the other 90% in a very low-cost S&P 500 index fund. He named Vanguard as an example of the kind of fund he meant.

That is the whole design. Two holdings, one of them a broad index fund, the other the shortest, dullest end of the government bond market. It is worth pausing on who wrote it: the person most associated with picking individual businesses told the trustee of his own family's money not to pick anything at all.

People often confuse this with Berkshire's 13F portfolio — the list of public equities Berkshire owns, which changes every quarter and is a different subject entirely. The 90/10 is a personal allocation instruction. It is also the version that keeps being searched for, because it is the one a private investor could actually copy.

The allocation

Sleeve	Weight	What it is
S&P 500 index fund	90%	Broad ownership of large US companies, at minimum cost
Short-term government bonds	10%	Money that will still be there when equities are down

No international allocation. No gold. No bonds of any real duration. No rebalancing rule stated. The instruction is remarkable mostly for what it leaves out.

Why each piece is there

The 90% is a bet that owning a broad slice of American business, held indefinitely and cheaply, beats trying to be clever — an argument Buffett has made publicly for decades, most visibly in the ten-year wager he made against a basket of hedge funds. The word doing the work is "low-cost": the case rests on fees not eating the return.

The 10% is not there for return. It is there so that the household has money it does not have to sell equities to raise. That is the whole job of the sleeve. Short-term government bonds move very little when interest rates move, which is precisely why they are chosen over longer-dated bonds: the point is availability, not yield.

Read that way, 90/10 is not an asset allocation in the usual sense. It is one growth engine plus a small reservoir that keeps the owner from being forced to touch the engine at the wrong time.

What this portfolio is built to survive — and what it is not

It is built to survive time. A long equity holding period, no manager risk, no strategy drift, almost no cost.

It is not built to survive a bad decade in US equities with a short horizon. A 90% equity weight means the portfolio does roughly what the S&P 500 does, including the parts nobody enjoys. Anyone copying it is accepting equity drawdowns almost in full, cushioned only by a 10% sleeve that will not move the total much.

The context matters too. Buffett was describing money left to a beneficiary with no need to spend it soon, alongside a great deal of other wealth. The same instruction given to someone who will need the capital in four years is a different instruction, even though the numbers look identical.

How it compares to other famous portfolios

Against the Permanent Portfolio (25% each in stocks, long-term Treasuries, cash and gold), the Buffett allocation is the opposite temperament: it concentrates on one engine instead of spreading across four economic states. Expect more growth in expansions and far more pain in equity bear markets.

Against the All Weather portfolio (30% stocks, 55% Treasuries split between long and intermediate, 15% gold and commodities), the contrast is starker still. All Weather is designed so that no single environment can dominate the outcome. 90/10 lets one environment dominate completely, on purpose.

Against a classic 60/40, the Buffett version simply takes more equity risk and holds shorter bonds. The 40% in a 60/40 usually includes duration, which is meant to rise when equities fall. Buffett's 10% is not trying to do that job at all.

When it underperforms

• Prolonged equity bear markets. With 90% in one asset class, there is nowhere to hide. The cash sleeve funds spending; it does not offset losses.
• When the money is needed soon. A drawdown that is survivable over twenty years is not survivable over three.
• Currency and concentration. It is entirely US large-cap equity. An investor whose liabilities are in euros carries a currency exposure the design never mentions.
• When "low-cost" is not honoured. The entire argument depends on fees being negligible. Implemented through an expensive wrapper, the logic weakens considerably.

Implementation notes

The design translates into two fund categories: a broad, low-expense S&P 500 tracker and a short-dated government bond fund or money-market instrument. European investors typically use UCITS versions, which changes the tax and domicile picture and sometimes the tracking cost — worth checking before assuming equivalence.

One detail people miss: Buffett specified an S&P 500 fund, not a total-market or all-world fund. Those are different exposures, and swapping them is a decision, not a detail.

Rebalancing

The original instruction sets no rebalancing rule. In practice a 90/10 mix drifts toward equities as they compound, so investors who copy it usually add a rule of their own — a calendar check once a year, or a band that triggers when the cash sleeve strays a few points from target. Adding a rule is sensible; it is also an addition, and it should be recognised as one rather than presented as part of the original.

Who it fits, and who it does not

It fits an investor with a long horizon, no need to draw on the capital soon, and the temperament to watch a portfolio fall a long way without selling. That last condition is the binding one, and it is not knowable in advance — most people discover their tolerance during a decline, not before.

It does not fit someone near the point of spending the money, someone who needs stable income, or someone who will not be able to leave it alone.

Frequently Asked Questions

What exactly is the Warren Buffett 90/10 portfolio?

Ninety percent in a very low-cost S&P 500 index fund and ten percent in short-term government bonds — the instruction Buffett described in his 2013 Berkshire Hathaway shareholder letter for the trust benefiting his wife.

Is the 90/10 the same as Berkshire Hathaway's portfolio?

No. Berkshire's equity holdings are a separate, actively managed list disclosed quarterly. The 90/10 is a personal allocation instruction for cash left in trust.

Why short-term bonds rather than longer-dated ones?

Because the sleeve exists to be spendable, not to earn. Short maturities barely move when interest rates change, so the money is reliably there.

Does it work outside the United States?

The mechanics transfer; the exposures do not. A non-US investor copying it holds US large-cap equity and US government debt, with the currency risk that implies. That may be acceptable, but it is a choice being made, not avoided.

How often should it be rebalanced?

The instruction does not say. Investors who adopt it commonly review once a year or when weights drift beyond a set band.

Is this investment advice?

No. This page explains a published allocation and the reasoning usually given for it. It is not a recommendation, and nothing here accounts for your circumstances, horizon or tax position.

Bottom Line

The Buffett 90/10 is famous because of who wrote it and how little it contains. Its logic is not that markets are easy — it is that costs and behaviour are the two variables an ordinary investor actually controls, and that a simple structure protects both.

What it is not is a low-risk portfolio. Ninety percent in equities behaves like ninety percent in equities. The design assumes a horizon long enough for that to be an advantage rather than a problem, and it assumes an owner who will not interrupt it. Strip either assumption away and the same two lines describe a very different experience.

Educational content, not investment advice. Past behaviour of any allocation does not predict future results.