The cautious version

Who this is for

Every other portfolio on this site fell by about half at some point since 1990. If you would have sold partway through that, you need a smaller fall, and the only way to buy one is to hold fewer stocks. This page gives two versions, one for a fall of about 40% and one for about 30%, each measured as a whole rather than assembled from rules of thumb. The labels come from what each version did since 1990. What each did across 1929 to 1932 is printed beside it every time, because on that history nothing on this site falls less than about 42% unless it holds under 38 points of stocks.

Too close to call The smaller fall is measured. Whether the trend line and the lean earn their fees inside it is not.

For a fall of about 40%

For a fall of about 30%

How far each fell

On the fund-based history from 1990 to 2026 the −40% version's worst fall was −18.1%, and it took 26 months to recover; across 1929 to 1932 its worst fall was −53.8%, with 74 months under water. The −30% version's worst fall was −15.8% since 1990, with 31 months under water, and −41.7% across 1929 to 1932. On every history that starts in 1934 or later the two fell 27% and 19% at worst. Portfolio three fell 50% since 1990 and 83% from 1929.

So the labels describe the past sixty years only. The 1929 crash put every mix of stocks and bonds between 12 and 33 points deeper than any modern fall. A reader who wants to be safe against that year needs fewer than 38 points of stocks, which is less than either version holds, and I have not scored a mix that small.

In 2022, the year stocks and bonds fell together, the −40% version returned −19.3% and the −30% version −18.2%. Bonds did not protect anyone that year: a portfolio holding only Treasuries returned −25.1%. Hold them for the smaller fall over a whole cycle, not as insurance against any one year.

What you give up

Return. Over 1929 to 2025 the −40% version grew 8.3% a year and the −30% version 7.4% a year, against 10.2% for an all-stock fund and 11.1% for portfolio three. Measured as a yearly gap against the all-stock fund, that is −2.1 and −3.2 points a year at the returns stocks actually paid, which were about 7.75 points a year over bonds.

Today the gap is smaller. Ten-year TIPS pay 2.44% above inflation, and most forecasters expect stocks to beat that by between nothing and 1.5 points a year. At nothing, holding half your money in TIPS costs nothing. At 1.5 points it costs the −40% version about 0.75 points a year and the −30% version about 0.9. At 3 points it costs about 1.5 and 1.9. You pick the row you believe. The one decade the cautious versions won outright was 1999 to 2009, when the −40% version ran +2.5 points a year ahead of portfolio three and the −30% version 3.1.

Against a plain mix of 60% stocks and 40% Treasuries with no lean and no trend, the −40% version earned +1.43 points a year more over 1990 to 2026 with a shallower worst fall. The −30% version's gap against the same mix is too small to tell from luck.

Whether the trend line earns its place here

Holding the stock share fixed and swapping RSST for plain stock funds, the trend line added +0.92 points a year to the −40% version and +0.68 to the −30% version over 96 years, and both versions fell less than their plain twins on every history. Since 2009 the gap has been too small to tell from zero, as it has been for portfolio three. At the trend returns I would expect from here it is worth about 0.3 to 0.4 points a year, and it stops paying for itself if trend following earns less than about 1.2 points a year before costs.

If you want none of it, the plain twins are the same stock funds and SCHP with no RSST: 51% stocks and 49% TIPS fell −21.6% since 1990 and 55.8% across 1929 to 1932; 38% stocks and 62% TIPS fell 18.4% and 43.6%.

Where to hold it

SCHP goes in a traditional IRA or 401(k) first, then a Roth, and in a taxable account last. Bond income is taxed as ordinary income, and TIPS are taxed each year on inflation you have not been paid yet. Half or more of the portfolio in TIPS is more than most people's sheltered accounts can hold, so some of it will sit in a taxable account. The which-account tool ranks the rest of the funds, and RSST still goes in a sheltered account before any of them.

What is not known

No TIPS existed before 2003, so on both long histories a plain ten-year Treasury stands in for them. That makes 1973 to 1974 and 1977 to 1981 read worse than a TIPS portfolio would have done, by an amount no data can measure; across 1977 to 1981 the two versions gained 36% and 24% while portfolio three gained 86%. On the years since 2003 where both can be run, TIPS earned 0.3 to 0.4 points a year more and fell a little deeper at the worst, and nothing in that window separates the two. The case for TIPS over plain Treasuries is the 2.44% real yield you can lock in today, not the history.

These are constructions built from index and futures series, not fund records. The trend line's measured help predates 2009. Every figure is before tax, and a large bond line in a taxable account is taxed harder than stocks.