Two. The same funds, held well

Why each piece is there

The holdings do the same job they did on the previous page. Four things change about how they are held.

Hold the cheapest version of each fund, where cheapest means the fee minus what the fund earns lending its shares out. That reordering is real. IEMG charges 0.09% and costs −0.009% once lending is counted; VWO charges 0.06% and costs 0.017%. The dearer fund is the cheaper one to own. SPY costs 0.0945% a year against 0.018% to 0.029% for every other S&P 500 fund audited, because its 1993 trust bars it from lending anything and parks dividends in an account paying no interest.

Put each fund in the account where it is taxed least, if you have more than one. With a single account, skip this step and lose nothing.

Prefer an exchange-traded fund to an old-style mutual fund. In 2025, 7% of exchange-traded funds paid out a capital gain against 52% of mutual funds. VOO, VFIAX, VTI and VTSAX distributed none at all across 44 fund-years; two long-standing active funds, AGTHX and FCNTX, distributed 6.62% and 7.01% of assets.

Do not trade. Rebalance about once a year, to a band rather than an exact target, inside a sheltered account where a trade costs no tax. Rebalancing keeps your mix from drifting. It is not a way of making money, and whether it costs or earns you a little depends on the period you measure.

What it is compared against, and what the comparison showed

The comparison here is against whatever you would otherwise have held, which makes it the only one on this site that depends on you. Assembled from fund filings and tax law rather than from a backtest, the whole package runs from about 0.04% to about 2.70% a year. Two things decide where you land inside that: how expensive the funds you own now are, and whether you have more than one kind of account.

The part that is certain and applies to everyone is the fee. Cutting a fund fee is contractual, it is the largest single line, and it arrives in months. Moving out of the fund you would otherwise have held saves about 0.49% a year, plausibly 0.40% to 0.59%. Against an index fund it saves nothing.

There is a central figure, and it comes with conditions. About 1.09% a year is what the package was worth to a modelled investor leaving an expensive active mutual fund, with a taxable account, more than one kind of account, and contributions continuing. It is a sum of separate lines, and several cannot be true for the same person at once. Roughly 0.23 points is for holding an exchange-traded fund rather than an active mutual fund, worth nothing to someone already in index funds. Roughly 0.30 points is for selling losers to offset gains, worth close to nothing to someone who takes the advice above and does not sell. About 0.05 points is for telling your broker which shares to sell first, a judgement rather than a measurement, and one that cannot hold alongside never selling.

Two smaller lines, honestly sized. Account ordering is worth about 0.02% to 0.07% a year against a plan you could actually execute, and turns negative if your only option is a captive employer plan. Never selling in a taxable account is worth about 0.14%.

Outside evidence points the same way. Sialm and Zhang found that US stock funds cost their taxable holders 1.12% a year in tax on average over 1990 to 2012, roughly the same as the fee, and that tax varies about three times as much between funds as fees do. The gap between the most and least tax-efficient fifth ran to 2.10 points a year, with nothing given up before tax.

How sure we are

Settled, in direction. Fees are set by contract, account treatment by statute, and the exchange-traded structure’s tax behaviour by a rule that has held for decades, so none of it needs a return forecast to be right. The size is a different question, and it depends entirely on what you hold now.

What it feels like when it is losing

It falls exactly as far as portfolio one, because it holds the same things. US stocks have fallen 83.7% from a peak. A cheap global index that borrows nothing fell 52.7% at its worst over the 427 months we tested, and in the ten years to February 2009 US stocks lost 2.55% a year.

The discomfort particular to this page is quieter. If you already hold cheap index funds in sensible accounts and rarely trade, this portfolio is worth nothing to you, and reading a page of savings while collecting none of them is its own kind of losing. If you hold something expensive in a taxable account with a large gain, moving may cost more in tax today than the fee saves for years. Expect, too, to watch a cheaper fund appear and do nothing, because chasing a 0.01% difference across a taxable account turns a certain saving into a certain tax bill.

What would make us drop it

The fee gap closing, which would happen if the funds people actually hold got cheap. A change to the rule that lets an exchange-traded fund hand back shares without realising a gain, which would remove the largest line in the tax comparison. Lending income shifting enough to reorder which fund is cheapest, which is why every figure here carries a read date.

The decision that matters more than this page

How much of your money is in stocks at all outweighs everything above. Moving a portfolio from 60% stocks to 90% was worth 1.27 points a year, against 4.85 points of variation from one year to the next. That is more than every lean on this site combined. Choosing 60/40 is itself a forecast: it says stocks will do better than bonds by only about 1.2 points a year. One country, 1963 to 2025, with a modelled ten-year government bond, and about 24 years of your own results before you could be 90% confident about the step.