Is my ETF portfolio diversified enough? Count independent bets, not funds

A portfolio is diversified by how its holdings move, not by how many it has. From January 2006 to April 2026, in euros, a mix of four equity ETFs behaved like 1.2 independent bets, barely more than a single World fund. A 60/40 mix with high-quality government bonds reached 1.4, and a six-fund mix across five asset classes 2.1.

The measure is a useful check on whether extra funds add a different source of return, or only more of the same risk; it says nothing on its own about which mix will grow wealth fastest.

Key takeaways

1. Fund count overstates diversification. Four equity ETFs had 1.2 independent bets and the same worst drawdown as MSCI World alone, about 49%.
2. Credit funds add less than they appear to. Adding 20% euro corporate and 20% high-yield bonds to world equities gave 1.2 bets, because both bond types fell with equities in most bad months.
3. Different drivers add bets. Mixes with government bonds, gold, inflation-linked bonds or cash scored 1.4 to 2.6 and lost between 7.7% and 29.0% at worst, against 49.0% for world equities.
4. The count moved with the regime. Since 2022, government bonds have fallen with equities more often, and the 60/40 mix's score dropped from 1.4 to 1.2.

Why counting funds misleads

Funds are wrappers. What diversifies a portfolio is the number of different things that drive its returns: company earnings, interest rates, inflation, the price of gold. Two funds that track the same driver add the cost and effort of two positions, and the risk of one.

Our falling-months study shows the problem in the months that matter: in world equities' 24 worst months since 2006, every equity region we tracked fell in at least 23.

A measure that counts drivers

The diversification ratio compares the risk of the parts with the risk of the whole:

Diversification ratio = weighted average volatility of the holdings ÷ volatility of the portfolio

If every holding moves in lockstep, the two are equal and the ratio is 1. The less the holdings move together, the more their swings offset, and the higher the ratio. Choueifaty and Coignard (2008) proposed it, and Choueifaty, Froidure and Reynier (2013) interpret its square as the effective number of independent risk factors in a portfolio. We call that the number of independent bets. A score of 2 means the portfolio's swings offset as much as two equally risky, uncorrelated assets held in equal parts would.

We also report how much of each portfolio's variance came from its equity holdings, using a standard risk-contribution split.

Seven portfolios, 2006 to 2026

We built seven hypothetical mixes from index returns in euros, rebalanced monthly without costs, from January 2006 to April 2026. They are examples of common structures, not recommendations.

Source: Copio calculation on index levels in euros; results by portfolio (CSV).

What the table shows:

• The four equity ETFs are one bet. Their worst drawdown was 48.7%, against 49.0% for MSCI World alone, and the worst 12 months were deeper (-39.4% against -37.6%). Adding regions changed which countries led, not the size of the fall.
• Credit cut volatility by dilution, not by offset. Swapping 40% of equities for corporate and high-yield bonds lowered volatility to 9.7%, but the bonds fell in many of the same months, so the bet count stayed at 1.2.
• Gold and government bonds added the most bets. Moving 10% from government bonds to gold took the 60/40 mix from 1.4 to 1.7 bets; the four equal quarters reached 2.6.

Diversification is not the same as growth

The scores rank how much the holdings offset each other. They do not rank which mix compounded wealth fastest, and this window flatters some assets. Gold returned 11.4% a year in euros from 2006 to April 2026, more than MSCI World's 8.7%, while EUR-hedged government bonds returned 1.7% and euro cash 1.0%. That is why the 60/30/10 mix grew faster than the 60/40 mix here. In the 24 years gold needed to regain its 1983 peak, the same comparison would have looked very different.

The four equal quarters show the trade-off most clearly: the lowest drawdown, 7.7%, and 6.1% a year, against 8.7% for world equities. Over 20 years, that gap compounds to a much smaller ending balance, and half the portfolio sat in assets with low expected returns. A mix with more independent bets lowers the risk of a forced sale at the bottom; the price is usually some expected return, and the right balance depends on spending needs, horizon and debts.

The count changes with the regime

Source: Copio calculation; results by portfolio (CSV).

From 2006 to 2021, equities accounted for 101% of the 60/40 mix's variance: government bonds slightly offset equity swings, so their own contribution was negative. From 2022 to April 2026, equities accounted for 85%, and bonds added risk of their own as they fell with equities in rate-driven sell-offs. The 60/40 mix's bet count fell from 1.4 to 1.2, the 60/30/10 mix's from 1.8 to 1.5, and the four equal quarters' from 2.6 to 2.2.

Two lessons follow. A mix that looks well diversified on one period's correlations can lose part of that benefit when the economic regime changes. And a high share of equity risk in a balanced fund is normal: in all these mixes except the four quarters, most of the portfolio's swings came from its equities, whatever the capital split. Our risk-contribution guide shows why a 60/40 split does not mean 60/40 risk.

Checks you can run on your own portfolio

1. Group your funds by driver before counting them. Equities from any region are one group; high yield sits closer to equities than to government bonds; gold, government bonds and cash are separate groups.
2. Estimate your own ratio. Take each fund's volatility from its factsheet or key information document, compute the weighted average, and compare it with your portfolio's volatility over the same period. A ratio close to 1 means one bet.
3. Check the equity share of risk, not of capital. If equities provide most of your swings, your crash exposure is close to your equity exposure, whatever the number of bond funds.
4. Test the period. Recalculate with data from 2022 onwards. If your bet count falls sharply, your diversification depended on the stock-bond relationship of 2006 to 2021.
5. Weigh the cost of each extra bet. Ask what the diversifying asset is expected to earn after costs and inflation, not only how it correlated in the past.

Limits of this analysis

• The mixes use index returns without fund costs, taxes or trading spreads, and assume monthly rebalancing.
• Twenty years is one sample with one financial crisis, one pandemic and one inflation shock; gold's strong run flatters the mixes that include it.
• The diversification ratio is built from volatilities and correlations. It does not capture rare extreme losses, liquidity or the risk of selling at the wrong time.
• The squared ratio is an interpretation of one statistic, not a count of economic factors, and it can rise without limit as correlations become negative.
• None of these portfolios is a recommended allocation.

How can copio help you examine your own allocation?

Copio is in private beta across Europe, with early access by application. It looks through your funds into asset classes and places each on a return-correlation map as a main driver, a true diversifier, dead weight or a shock absorber, so you can see whether your holdings differ by driver or only by label.

A correlation matrix shows which holdings move together, and a crash test shows how a mix like yours would have fared in past crises. Copio provides educational analysis, not personalised advice; it does not place trades or choose an allocation for you. The free portfolio checker is a first step.

This article is published for information and educational purposes only. It does not constitute personalised investment advice. Past performance is not a reliable indicator of future performance.