A 60/40 Portfolio Is Not One Strategy

Two portfolios arrive on the same page of the same board pack. Both report sixty per cent equities and forty per cent fixed income. The committee compares them on fee, on manager tenure, and on three-year performance, and moves on.

They are not the same portfolio. They may not even be the same kind of portfolio. One might hold domestic mega-caps against Treasuries. The other might hold global equities against private credit, emerging-market debt, inflation-linked bonds and a duration overlay. The headline is identical. Almost nothing underneath it is.

Start with what is actually on sale. These are five live funds, all currently marketed under balanced, moderate or tactical allocation labels, and between them they hold $545bn. Each is shown at its own portfolio date, because they do not share one.

The two funds circled are the sharper finding. They do not merely sit at different points on a spectrum. They carry the identical category tag, and they sit 18.5 points apart on equity with unrelated strategies underneath. One runs trend and macro overlays. The other is a levered fund of funds built around inflation-linked bonds and real assets.

That second fund reports a cash position of minus 21.5 per cent. It is borrowing, through futures and repo, and the borrowing is disclosed in the filings. It is not disclosed by the words on the front of the fund. A plain-language allocation name gives no indication that leverage is present at all.

The obvious objection is that those five funds are not really 60/40 vehicles, so of course they differ. Fair. So we removed the variable.

We built five portfolios that are precisely sixty per cent return-seeking and forty per cent risk-reducing, to the decimal, and changed only what sits inside each sleeve. Any divergence that survives is composition and nothing else.

It survives. Across the five, expected return spans 146 basis points and volatility spans 280 basis points, with no leverage and no exotic instruments anywhere in them. The concentration measure runs from 0.24 to 0.85 on a scale where 1.0 is a single holding, and the effective number of independent bets runs from 1.18 to 4.17.

The classic domestic construction scores 1.18. It has a four-line allocation schedule and it behaves as a two-asset bet. The pension-style construction, built from the same sixty and forty, scores 4.17. Both are ours and both are re-derived.

The live funds sit across the same range and reach further at the top. Those scores are provisional pending independent re-derivation, so they are reported here as direction rather than as settled figures. On that basis two things stand out. The longest-running balanced fund in the set sits near the bottom, running one dominant exposure with a bond position attached. And two funds less than a point apart on the headline split, at 65.5 and 66.3 per cent equity, are materially different underneath, because one puts roughly a third of its equity sleeve outside its home market and the other does not.

Here is the part that a paper making this argument is tempted to bury.

Every one of these portfolios, constructed and live, is still dominated by a single factor. Growth beta explains the large majority of the variance in all ten. The creative ones are less dominated than the plain ones, but none of them escapes. If you were hoping that a cleverly built 60/40 becomes a genuinely balanced set of risks, it does not.

That does not weaken the argument. It sharpens it. The differentiation between these portfolios is real but it lives at the margins: in tail loss, in how much inflation protection is actually present, and in whether illiquidity has been taken deliberately or by accident. Those margins are exactly what the headline label cannot show you, and exactly what determines behaviour in the year the portfolio is tested.

The one portfolio in the set that scores best on risk efficiency is the one deliberately built against something other than growth beta. That is the tell. A book holding that profile under a 60/40 label made a choice. A book holding the concentrated profile under the same label may simply have inherited a default that nobody has interrogated since it was set. The label cannot distinguish between those two boards. It reports the same number for both.

Everything above was computed on one house's capital market assumptions. That is standard practice, and it is also a choice, and the choice is not neutral.

We hold the published long-term assumptions of eight houses in one place, normalised to a common structure. So we took the simplest possible portfolio, sixty per cent US equity against forty per cent US core bonds, held it completely fixed, and priced it on each house's own numbers. Same weights, same currency, same ten-year horizon, same vintage.

And the disagreement is not a simple optimism dial, which would at least be easy to adjust for. It is a disagreement about shape.

They differ by nearly three times on whether the sixty is worth having at all. Run an optimiser on the low number and it will tell you to hold less equity. Run it on the high number and it will tell you to hold more. Both are defensible published views from serious institutions.

One further detail, because it is the sharpest of the lot. One of these houses publishes both an arithmetic and a compound expected return for the same US equity line, and the two differ by 124 basis points. That is a choice about which of two correct numbers to use, and on its own it is larger than the 99 basis points of disagreement between all six houses, and comparable to the 146 basis points of difference between five genuinely different portfolios.

Not “60/40 is dead”. That argument has been made many times and it is not what this shows. What this shows is that the label is a classification, and a classification is not a risk statement. Three questions follow, and they are answerable in an afternoon.

• What is actually inside the sleeves?

  Not the asset-class percentages. The geography, the credit quality, the duration, and whether anything in the risk-reducing forty is behaving like equity when it matters.

• How many independent bets is the book really running?

  If the answer is close to one, the allocation schedule is decoration. That is a governance finding, not a performance finding, and it does not appear anywhere in a standard report.

• Whose assumptions produced the expected return you are planning against?

  And what happens to it under someone else's? If the number moves by a hundred basis points and nobody in the room knew it could, the planning conversation was never really about the portfolio.

None of these require a new mandate, a new manager or a new system. They require decomposing what is already there.