The three-fund portfolio and why it is hard to beat

The three-fund portfolio is exactly what it says: one fund that holds the whole U.S. stock market, one that holds the whole stock market outside the U.S., and one that holds the U.S. investment-grade bond market. Nothing is picked, nothing is timed, nothing is predicted. The idea was popularised by Taylor Larimore, a co-founder of the Bogleheads community whom John Bogle himself called "the King of the Bogleheads," and it has a book-length treatment with Bogle's name on the cover. This page is why something that simple is so hard to beat — and what it cannot do, which is also important.

What three funds actually hold

"Total" means the whole market, not the big names: large, mid-sized and small companies alike, at the weights the market gives them. Between the U.S. fund and the international fund, you own nearly every company a global index tracks. MSCI's all-country index held 2,414 companies across 23 developed and 24 emerging markets at the end of September 2026, and its country weights say where the world's stock-market value sits:

A U.S. fund plus an international fund, held roughly in those proportions, is as close to "own the world's companies" as a person can get. The bond fund is the part that does not move with them — the ballast, and the part you draw on when stocks are down.

What it costs

Almost nothing, and that is the point rather than a detail. Vanguard's own fact sheets, read in October 2026, list expense ratios of 0.03% for the total U.S. stock ETF (VTI), 0.04% for the mutual-fund version (VTSAX), 0.03% for the total bond ETF (BND) and 0.05% for the total international stock ETF (VXUS). On $10,000, that is three to five dollars a year. Vanguard renamed the U.S. fund "Vanguard Morningstar Total Stock Market" in 2026 and it now tracks a Morningstar index; the holdings are the same idea.

Cost is the one input you control completely, and it is the one that predicts best. Morningstar's 2016 study of its own fund database concluded that "the expense ratio is the most proven predictor of future fund returns": the cheapest fifth of U.S. equity funds succeeded — survived and beat their category — 62% of the time over 2010–2015, against 20% for the most expensive fifth. Three times as likely, for the single act of paying less.

Why it is hard to beat

Two reasons, one of them arithmetic. William Sharpe's 1991 argument is that every dollar in the market is held either passively or actively; the passive dollars earn the market return by definition, so the active dollars as a group must earn the same before costs and less after them, in any period. The three-fund portfolio is the passive dollar. Everyone trying to beat it is, in aggregate, paying for the privilege of trying. The index-funds page has the long-horizon numbers: over twenty years, about nine in ten professional large-cap managers trailed the S&P 500.

The second reason is older. Harry Markowitz's 1952 paper is the start of the idea that an investor wants the maximum expected return for a chosen level of risk, and that spreading money across things that do not move together gets more return per unit of risk than concentrating it. Before Markowitz, as a recent retrospective puts it, "diversification and risk were fuzzy concepts." The three-fund portfolio is the plainest possible implementation of that idea: own everything, so that no single company's failure — and most companies do fail to beat Treasury bills — can matter much.

The part that is about you

A simple portfolio is also a portfolio you are less likely to wreck. Vanguard's "Advisor's Alpha" research followed 58,168 self-directed IRA investors from 2008 to 2012. Those who traded in and out of their holdings trailed the target-date fund they could have simply held by 150 basis points a year — 1.5% — against 19 basis points for those who left things alone. The researchers put behavioural coaching at the top of the list of things an adviser is worth, at an estimated 1% to 2% a year. The three-fund portfolio's real feature is that there is nothing in it to fiddle with.

What it cannot do

It cannot avoid drawdowns. The stock half will fall exactly as far as the stock market falls; a 60% stock allocation does not mean a 60% share of the pain, it means the whole of the stock market's pain on 60% of the money. The drawdown page has the recovery arithmetic. It cannot tell you when to put money in — the lump-sum page has that evidence. And it cannot make the one decision that matters, which is the split.

The one decision it leaves you

Everything above is settled by evidence. The stock-to-bond split is not; it depends on how long the money can stay invested, what you will need to withdraw and when, and how much of a decline you can watch without selling — the last of which nobody knows about themselves until they have done it. Larimore's own writing walks through it; this page does not pick a number for you, because a number picked for a stranger is a guess dressed as advice.

What this has to do with a futures desk: nothing, except candour. The desk's systems are not an alternative to any of this. At most they are a small sidecar, run with money whose loss would not touch the plan above, by someone who has read this page and the shelf it sits on.

• Rebalancing — how often, and what it is really for
• Dollar-cost averaging vs lump sum: the evidence
• Taxes: what the account decides before the market does

← Long-term portfolios · All of Learn