The Three-Fund Portfolio

Three funds, one of which does something completely different from the other two. On $50,000, choosing the ratio badly costs about $4,500 in a bad year, and choosing the wrong country split costs about a tenth of that.

This piece assumes you have read what an index fund actually is and what it costs, because the fee is a bigger decision than anything here. What follows is how to build the thing and buy it, probably today.

The three holdings and the job of each

A total US market fund owns roughly every publicly traded US company, weighted by size. This is the engine, and it is the reason the portfolio grows at all.

A total international fund owns companies outside the US, usually across both developed and emerging markets. Its job is not extra return. Its job is that the decade the US market disappoints is not automatically your decade, and the honest description is insurance whose premium is paid in slightly lower expected return.

A total bond fund lends money to governments and companies and collects interest. It is not there to grow. It is there to fall less than shares do at the moment you might be tempted to sell, and in a good year it will feel like dead weight. That feeling is the price of the function.

Notice that the first two are variations on one thing and the third is a different instrument. That is why the bond share is the decision that moves the result, and the country split is the one people argue about.

Picking the ratio

Here is what three ratios do to the same $50,000 in two bad years. The assumptions are stated because the whole table depends on them: the US market falls 37%, international falls 44% and bonds gain 5%, which is approximately the shape of 2008; then the US falls 19.5%, international falls 16% and bonds fall 13%, approximately the shape of 2022.

Two findings, and the second one is the useful one.

The bond share is what moves the number. Going from 10% bonds to 40% bonds turns a 34.1% fall into a 21.0% one, which on $50,000 is $6,510 of difference in a single year. That is the lever, and it is the only one in this article large enough to be worth an evening.

And 2022 exposes the limit of the lever, which almost nobody mentions when recommending bonds. In 2022 bonds fell too, so the 40% bond portfolio lost 16.5% against the aggressive one's 18.2%. A difference of 1.7 percentage points. The protection bonds offer is real in a year when shares fall and rates do not rise, and close to absent in a year when both happen at once.

So choose the bond share on when you need the money, not on how bad a year feels. Money you will not touch for twenty-five years can carry a small bond share or none, because you have time to recover and bonds reduce the expected return. Money you will start drawing in eight years should not be 90% shares, because a 34% fall in year seven is not a paper loss for you. The 60/20/20 row is a reasonable default for someone a couple of decades out who has never lived through a crash with real money in the market.

On the country split: the difference between 20% and 40% international, holding bonds constant, is about $2,000 on $50,000 in a 2008-shaped year. It has consumed more argument than any other question in index investing and it is worth less than checking your expense ratio once.

Where should each fund live?

The same three funds, in different accounts, produce different after-tax results. Two rules cover most of it.

Bonds go in the tax-advantaged account where you have the choice. Bond interest is taxed as ordinary income in the year you receive it, so holding a bond fund in a taxable brokerage account creates a tax bill every year on money you are not spending. Inside a 401(k) or IRA that interest compounds untaxed.

The 401(k) menu decides what is possible, so build around it rather than against it. A typical arrangement: bonds and whichever broad equity fund is cheapest inside the 401(k), then the funds the plan lacks in an IRA, then anything left over in taxable. For 2026 the IRA limit is $7,500, or $8,600 from 50, and the 401(k) elective deferral limit is $24,500.

One warning about holding the portfolio across several accounts: the percentages are for the total, not for each account. A 401(k) that is 100% bonds is not a bond portfolio if the IRA beside it is 100% shares and three times the size. Keep one sheet with all three accounts on it, or the ratio you chose is not the ratio you own.

Buying it, step by step

About twenty minutes, plus one settlement day if you are moving money in.

1. List what each account can buy, with the fees. Sort your 401(k) menu by expense ratio, not by past return. Write down the cheapest broad equity fund, the cheapest international fund and the cheapest bond fund, with their numbers.
2. Choose the bond percentage. One number, decided on when you need the money. Write it down before you look at any prices.
3. Split the rest. 70/30 or 80/20 US to international. Pick one and do not revisit it.
4. Place the bond fund where it is not taxed annually, then fill the equity funds around it across your accounts.
5. Buy. A mutual fund order fills once after the close. An ETF fills during the day; a market order on a large fund is fine. The amount is whatever the fourth bucket of your spending plan can actually support, month after month, rather than a figure that assumes a quiet year.
6. Set the recurring contribution, and check the allocation applies to future contributions as well as to the existing balance. These are separate settings in most workplace plans, and setting only the first is the most common mistake in this whole process.
7. Put one rebalancing date in the calendar and close the tab.

Rebalancing is arithmetic rather than judgement, and here is what it looks like after a genuinely bad year. A 60/20/20 portfolio that started at $50,000 and lived through the 2008-shaped column now holds $18,900 of US, $5,600 of international and $10,500 of bonds, which is $35,000 in total and 70% in shares rather than 80%.

Restoring the target means moving $3,500 from bonds into the two equity funds. Which is the uncomfortable part, and the reason to decide the date in advance: rebalancing in March 2009 meant buying shares in the worst month in thirty years, with a straight face, because the calendar said so. That is the entire discipline, and nobody has ever enjoyed it.

What would a fourth fund add, honestly?

Usually cost and overlap. Four of the popular additions, with what they actually do:

Before adding anything, open your existing funds' holdings and find out what is already in them. Most fourth funds are bought to fix a gap that does not exist, and each one is another expense ratio and another line to rebalance for the next thirty years.

When this does not apply

If you have no cash reserve. Build one month of essential spending first. A portfolio you have to sell into a bad month to fix a boiler is not a portfolio, and sizing that cash properly comes before any of this.

If your plan offers a cheap target-date fund. It holds a similar mix, adjusts the bond share for you and rebalances itself, which removes the two steps most likely to be forgotten. If its expense ratio is within about 0.10% of what the three funds would cost you, take the simpler option and spend the evening elsewhere.

If the money has a date inside five years. A deposit, a wedding, a sabbatical. No ratio makes shares appropriate for money with a deadline, because the market does not consult your calendar. Cash or a matched-term CD is the answer, and that is a different article.

If you are already drawing on the portfolio. The question changes from ratio to withdrawal order and sequence risk, and a three-fund split designed for accumulation is not automatically right for someone spending from it.

If you know you will sell in a crash. Not "might". If you have done it before, the right portfolio is the one you will hold, which probably means a larger bond share than the arithmetic recommends. A 21% fall you sit through beats a 30% fall you sell into, and the difference between those two outcomes is far larger than any ratio on this page.