Ten ETFs Is Where Diversification Stops Working
A friend of mine — let's call him Rashid, though he could just as easily be Marc, Jean-Paul, or Minh — spent two years teaching himself investing and ended up with seventeen ETFs in a modest account. Global equity, S&P 500, Nasdaq, Europe, emerging markets, a few thematic bets, sector funds, bonds of every flavor, gold, commodities. Nothing wrong with any single piece, each one does exactly what it says on the label. The trouble is that none of the seventeen answers a simple question: what problem does this solve that something already in the portfolio doesn't solve just as well?
Open the hood and two things fall out. First, overlap: several funds ending up owning, in slightly different weights, the same twenty American mega-caps. A broad U.S. fund and a global one typically share around sixty percent of their weight, so buying both means paying twice for the same cake, cut with two different forks. Second, diminishing returns: past a point, adding another fund stops spreading risk and starts multiplying the number of tabs to check and the number of moments to sell in a panic.
That point sits closer to intuition than you'd expect. Thomas et al. (2018) studied eighty-eight of the largest U.S. pension plans, institutions running an average of seventy-five external managers each, and found that beyond roughly ten equally weighted strategies, active risk collapses toward zero. Fees paid per unit of genuine differentiation nearly triple, from nineteen to fifty-three basis points, going from one manager to ten. Committees with research budgets hit the same wall retail investors sense by instinct: somewhere around ten holdings, more stops meaning better.
Not all extra ETFs are equal, and the distinction matters more than the count. Ben-David et al. (2022) tracked over a thousand U.S. ETFs launched between 1993 and 2019, splitting them into broad, boring funds and narrow, exciting ones chasing whatever story was fashionable. The specialized funds charged nearly double the fee and underperformed their broad-based cousins by roughly thirty percent, risk-adjusted, over the five years after launch, typically because they listed just as the hype behind them had already peaked. That's the eighteenth-ETF problem: not redundant, but actively expensive, bought at the worst moment for the worst reason.
None of this is new. Evans et al. (1968) found that a portfolio of just ten individual stocks already carries roughly the risk of the broad market. ETFs, which each bundle hundreds or thousands of stocks already, get there faster still: one global equity fund alone covers ninety to ninety-five percent of the investable world. Add a second piece for bonds, a third for real assets like gold, and the basic architecture is done. What separates a beginner's two-fund portfolio from a professional's five-to-ten-fund one isn't more of the same, it's a handful of genuinely different return sources: value and momentum tilts, inflation-linked bonds alongside nominal ones, trend-following strategies built to hold up when stocks and bonds fall together, as they did in 2022.
Put those findings side by side, a pension fund managing billions and someone with a brokerage app, and the coincidence stops looking like coincidence. Both hit the same ceiling, independently, without comparing notes. That's not a rule someone invented to keep things simple. It looks more like a structural fact about how diversification runs out of things left to diversify.
References
Thomas, McKay & Shapiro (2018)
What Free Lunch? The Costs of Overdiversification
Ben-David, Franzoni, Kim & Moussawi (2022)
Competition for Attention in the ETF Space
Evans & Archer (1968)
Diversification and the Reduction of Dispersion: An Empirical Analysis
Key takeaways
- Overlap, not the raw count of funds, is often the real problem -- a broad U.S. ETF and a broad global one can share roughly sixty percent of their weight.
- A study of eighty-eight large U.S. pension plans found active risk collapsing past roughly ten equally weighted strategies, with fees per unit of differentiation nearly tripling.
- Specialized, thematic ETFs underperformed broad-based funds by about thirty percent risk-adjusted over five years, largely because they launch right as their story peaks.