What to hold
Four portfolios you could actually hold, what each piece is for, and how much of the clever stuff holds up against a cheap index fund.
| Number | Portfolio | Holdings | Fees a year |
|---|---|---|---|
| 1 | The whole market | VT 100% | 0.06% |
| 2 | The same fund, held in the cheapest form, in the account where it is taxed least, and never traded | VT 100% | 0.06% |
| 3 | A lean toward cheaper, smaller, more profitable companies | VTI 49%, VXUS 16%, VTV 15%, AVDV 10%, IDMO 5%, AVES 5% | 0.09% |
| 4 | The same, plus a holding that does not move with stocks | RSST 30%, VTI 19%, VXUS 16%, VTV 15%, AVDV 10%, IDMO 5%, AVES 5% | 0.38% |
Rows one and two hold the same fund. What separates them is how it is held, and how much that is worth depends entirely on what you hold now.
Fees are each fund’s published expense ratio, read from filings on 17 August 2026, weighted by the holdings above. On $10,000 that is $6 a year at the top and $38 at the bottom.
Whatever share of your money you keep in bonds or cash sits outside all four. That choice is yours, and it matters more than the difference between any two rows above.
What is certain
Fees are contractual. Which account a fund sits in is statute. Whether a fund is exchange-traded or an old-style mutual fund largely decides whether it hands you a tax bill you did not ask for: in 2025, 7% of exchange-traded funds paid out a capital gain against 52% of mutual funds. None of it is a forecast, and all of it shows up within months.
As a package, holding well 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. One account already full of cheap index funds puts you at the bottom of the range.
The fee is the part that is certain and applies to everyone. Cutting it is contractual, it is the largest single line, and it arrives in months. Someone leaving a typical active fund saves about 0.49% a year, plausibly 0.40% to 0.59%. Someone already in cheap index funds saves nothing.
What is a bet, and how big
Portfolio three leans the stock holdings toward companies that are cheap relative to what they own and earn. Against a cheap global index fund that lean came to 0.79% a year, plausibly 0.30% to 1.32%. Telling the gap from zero would take about thirty years of data, and cheap US stocks have spent 17.7 years trailing the US market by 54.3% without catching up.
Portfolio four adds a fund that borrows inside itself to hold managed futures on top of stocks. It is the largest bet here and the least settled. Read its page before you touch it.
What we could not tell
Against a cheap index fund, everything clever here nets out to somewhere around zero. The central figure is 0.054% a year, with a range from −0.92% to +0.83%, and it crosses zero, so we cannot pin down even the sign.
That is smaller than the lean above it, and the reason is worth knowing rather than hiding. The 0.79% is the lean on its own. The near-zero total is what remains after netting it against everything else we measured, and those other lines are small, point both ways, and change sign depending on how far back you look. Read the total as around zero rather than as a precise number.
Some things failed outright, and we say so: bitcoin as ballast, funds that promise to rise when stocks fall, and crash insurance. Rebalancing is its own case. It keeps your mix from drifting, which is the reason to do it, and whether it costs or earns you a little depends on the period you measure. Other things we could not measure at all. A 200-day timing rule on US stock data came out 0.74% a year ahead of a fixed blend holding the same average amount in shares, against a test that could only have seen a gap of 3.03% or wider. Measured against simply holding, the same rule lost 0.73% a year. Those are two different questions and the answers point opposite ways.
Where to go next
The portfolios page puts all four side by side with questions to help you pick. The funds page lists every fund we priced and what each costs after the income it earns lending its shares out. The evidence page has every idea we tested and how it went.