Four ways to hold it

One

A cheap global index fund

Confidence: Settled Arithmetic, a contract or a statute. The sign is known before you start.

Three funds. The whole US market, the developed markets outside it, and the emerging ones. You own a slice of nearly every listed company on earth, and the fee column above is what you pay for it.

Who it is for: everyone, as a starting position. It is also the yardstick. Every other portfolio here gets measured against a fund like this one, because a fund like this one is what you can buy instead.

What it is worth over itself is nothing, by construction, and that is the point of it. Its case is that the alternatives were priced and lost, which is a weaker claim than saying it is good and a far easier one to check. The costs behind it come out of fund contracts, so their sign is known before you start.

The split between the three regions is a research weight rather than a measured optimum, and this project holds no data that could tell one sensible split from another. Pick one and stop worrying about it.

What a bad decade costs you: it falls when markets fall, and the record has some very deep falls in it. It also means watching one part of what you own run away from another for a very long time. International stocks once spent a stretch 69.0% behind US stocks and had not caught up by the end of the record.

Two

The same three funds, held properly

Confidence: Settled Arithmetic, a contract or a statute. The sign is known before you start.

Nothing changes about what you own. What changes is the fee, which account each fund sits in, whether you hold an exchange-traded fund or an active one, which shares your broker sells when you sell, and how often you trade at all.

Against the portfolio a normal person is already holding, that package is worth about 1.09% a year, and you would know inside about 3.5 months whether you got it right. Every line of it comes out of a fund contract or a tax statute. Nobody has to be right about markets for it to pay.

Against a cheap index fund instead of against your own old holdings, the same work is worth 0.054% a year, which is too small to tell from zero. Two questions, two answers, and adding them is the mistake this project has made most often against itself.

Who it is for: anyone paying an expensive fund, or holding whatever they bought in whichever account they happened to open first. If you already own one cheap index fund in one account and never touch it, your number here sits close to zero, and there is nothing to sell you.

Most of the lines below need a condition you may not have. The first one needs nothing but an expensive fund to escape from. The rest want a taxable account, or gains to offset, or more than one kind of account to sort between. Where you have none of that, the line is worth zero to you and the total shrinks accordingly.

What a bad decade costs you: nothing the first portfolio does not also cost you, because this one owns the same three funds. What it can cost you is the trading. Reorganising a taxable account means selling, selling means a tax bill, and that bill comes off the top of everything in the table. The advantage of holding an exchange-traded fund rather than an active one is also closing on its own, because a great many active managers have now filed to run one.

Three

The same funds, leaned toward cheap companies

Confidence: Might The mechanism is real and the measurement is thin. Size it as if you could be wrong.

Two of the three index funds give up part of their weight to funds holding the cheaper half of the same market, in the US and in the developed markets outside it. Nothing is borrowed and nothing else moves.

These are the only two leans this project has priced whose case survived its own comparisons. Both funds deliver the exposure they advertise, both are cheap, and both trade slowly, which matters more than the fee once a tax bill is in the picture. The emerging market has the largest measured gap of the three regions and no fund on this shelf was good enough to buy it with, so that piece stays plain.

Who it is for: a reader who wants a lean and wants it sized by what was measured rather than by how strongly they feel about it. The price of admission is decades of wandering away from the index for tens of basis points a year of hoped-for gain.

Read the two rows below as pairs. In both of them the wander column is several times the worth column, and the second number is the one that decides how your account will feel.

How sure we are: exploratory is the highest status anything in this project has reached, and all it licenses is standing in for a real product in a later test. It does not say the lean works. A study would need about 30 years at 80% power, on error bars that allow for markets clustering of data before it could separate this gap from zero. Charge the foreign fund the shortfall it has actually run against its own model and the second row turns negative. Four foreign funds of that kind read the same way, and nobody here knows why.

What a bad decade costs you: long stretches behind the plain index fund are the normal behaviour of this portfolio rather than evidence against it. Cheap US stocks are 54.3% behind for 17.7 years, and they had not recovered when the record was read.

Four

The stacked proposal

Confidence: Can't tell The test could not have seen an effect this small. Neither a yes nor a no.

A reader's proposal, priced against the same shelf as everything else here. Eight funds: the two leans above, a momentum fund, an emerging value fund, and one fund that holds US stocks and a managed-futures strategy on the same dollar by borrowing inside itself. Total exposure comes to 132% of the money you put in.

The most interesting number in it has nothing to do with any strategy. Paying for an addition by borrowing rather than by selling something you already own is worth about 2.44 percentage points percentage points a year on an all-stock base, and that figure contains no property whatsoever of the thing being added. It is the largest single result this project has, and it is a statement about how you pay rather than about what you buy.

Who it is for: a reader who has decided that spreading the bet across different kinds of return matters more than keeping the thing simple, and who can hold an unfamiliar losing streak without selling it.

How sure we are: two of the priced lines below are exploratory, one holding is unresolved, and the borrowed piece was rejected on this project's own bar. The momentum fund is excluded by name by the fund audit, because it turns over more than its own value every year and the cost eats the case. Two of the leans sit at weights other than the ones they were priced at, so the published figures do not describe this portfolio. And the borrowing itself has never been shown to pay here.

What a bad decade costs you: this is the one that will feel worst. The borrowed piece has spent a stretch 59.9% behind the stocks it displaces, and it did not recover. 98% share of active variance of the portfolio's active risk sits inside one young fund from a small issuer, in a category where 52% share of funds of the funds filing at the start of the window had stopped filing by the end of it. None of that appears in a fee table.

The detail

The full detail of the last one

The portfolio writes a construction of this kind out in full: every holding, the account each one goes in on a stated set of assumptions, and an honest account of how bad it feels while it is losing. The shelf lists every fund this project has priced, most of which were read in order to rule them out.

Choosing

How to choose

Nothing on this page depends on what you earn, and nothing on it assumes which accounts you have. Where an account type decides an answer, it is named as a condition rather than issued as an instruction. If you have a workplace plan. If you have a taxable account. If you have gains to offset.

If you want the arithmetic run at your own tax rates and your own account sizes rather than at a stated example's, the placement tool does that, and it is the only form of that answer worth acting on. The waiting tool tells you how long a gap of a given size would take to show up in an account of yours.

Two comparisons are worth carrying away. Over a career, the one research result this data actually resolves is worth about as much as raising what you put in each year by 3.08% of your starting balance each year. And none of these four holds a bond, which is a position rather than an oversight and ought to be a deliberate one. Shifting the balance between stocks and bonds moves the answer further than every lean on this page put together: +1.27% a year against 4.85% of drift.

Sources

Where this comes from

The four constructions and every number attached to them were read from the research on 2026-08-17. Fee and tax figures come from fund filings and tax statute rather than from a backtest. Everything measured against an index fund carries a wander that was assumed rather than measured, so read every probability here as a best case.

• The recommended portfolio
• Structural and tax edges
• Where outperformance can come from
• The factor-product audit
• Factor persistence
• Why nothing here is promoted
• Capital efficiency and breadth
• The marginal value of trend

How sure we are sets out the four words this site uses for confidence and how long each claim would take to settle. Corrections lists everything published here that later turned out to be wrong.