“Diversification” Is a Fiction: the S&P 500 “Gained” 0.2% in September While 75% of Its Stocks Fell — A Concentration X-Ray of MPF North American Equity Funds

In September, the S&P 500 gained 0.2% on paper. If that is all you saw, you would think US equities sailed through the rate-hike regime unscathed.

It is an illusion. Of 499 index companies, 387 fell during the month — more than three in four. The typical S&P 500 stock lost nearly 6% on average. Recompute the index with equal weights and September was -4.4%: the worst single month since March, and the widest monthly outperformance of cap-weight over equal-weight since March 2020.

Meanwhile, MPF members are pouring money into “America”: GUM’s August report shows US equity funds have taken net inflows of more than HK$12 billion year to date — five straight months of buying — while Hong Kong equity funds saw net outflows of over HK$10 billion. The catch: the “America” they bought is really seven stocks.

The X-Ray: Promised vs Actual

The shelf label on a North American equity fund says “diversified exposure to the US market.” September’s data tears the label off:

Why didn’t the index fall? Two stocks: Meta added US$424 billion in market capitalisation during the month, Apple added US$183.16 billion (Barron’s). Two companies held up the entire index and masked a straight-line selloff in the other 490.

Concentration itself is the story. As of January 2026, the Magnificent Seven (Alphabet, Apple, Amazon, Meta, Microsoft, Nvidia, Tesla) accounted for roughly 34% of the S&P 500; add Broadcom, Berkshire Hathaway and Eli Lilly and the top ten reach 40% (ATB Wealth). For every dollar invested, 40 cents go directly into ten stocks; the remaining 60 cents are sprinkled across the other 490 companies. You are not buying the market — you are buying a technology-concentrated fund with a side of general stocks.

History offers an uncomfortable comparison. Since the index’s creation in 1957, the average weighting of its seven largest constituents has been about 17%; the previous concentration record was roughly 26%, set during the 1980 energy boom and again at the March 2000 dot-com peak (Seeking Alpha). Goldman Sachs puts the dot-com peak for the top ten at 27% — today’s 40% has sailed well past it. ATB’s historical review adds a warning: the last time the top ten exceeded 30%, in the 1950s and 1960s, the market suffered three separate bear markets of 20% or more.

Cumulative returns since the start of 2023 complete the picture: the index +95%, the Magnificent Seven +150%, the equal-weighted index just +50% (Seeking Alpha). The SPY’s glossy +13.37% year-to-date figure (Barchart) is almost entirely the Magnificent Seven’s work. As Morningstar put it: “Tech is hanging in there, and everything else is just falling quite honestly in a straight line.”

The Compounding Cost: Concentration’s Bill

Bill one: the single-month illusion tax. MPF Ratings’ September data puts the average balance at HK$343,242. Fully invested in a cap-weighted North American equity fund, the headline “+0.2%” equals roughly +HK$686 — barely noticeable. But the typical constituent’s true fate (the equal-weighted -4.4%) means a loss of about HK$15,100. Same “US market,” two destinies HK$15,789 apart. What the index never told you is what your money actually went through.

Bill two: the arithmetic of a concentration reversal. The Magnificent Seven are 34% of the index, so the mechanical pass-through is brutal: a 20% pullback in the seven drags the index down about 6.8%, almost unaided. If 30% of your average balance sits in North American equity funds (HK$102,973), a single such event vaporises roughly HK$7,000 — about a month’s contributions. That is not a forecast; it is leverage built into the weight structure.

Bill three: the long-run structural gap. An illustrative calculation: HK$5,000 a month for 30 years compounds to about HK$6.10 million at 7% a year, HK$5.02 million at 6%, HK$4.16 million at 5%. One percentage point of annual return, sustained over 30 years, is worth HK$1.08 million — roughly 3.1 times the average balance. With 40% of index weight tied to ten stocks, any single giant’s re-rating (see September’s demonstration: Citi cutting Moderna from Hold to Sell, the stock down 5.4% in a day) is deducted directly from your compounding engine.

Bill four: the momentum-chaser’s timing tax. The SPY’s +13.37% year-to-date pulled in HK$12 billion of chase money — but September proves that money bought prices already propped up by the Magnificent Seven, while the 387 fallen stocks held the next cycle’s discounts. MPF’s forward-pricing mechanism (T+1/T+2 unknown-price dealing) makes chasing more expensive still: by the time you see September’s data, October’s dealing price is a different world.

The Breakout Plan

First, read the label honestly. A North American equity fund is approximately a 34% Magnificent-Seven technology concentration fund plus 490 side dishes. Even the DIS Core Accumulation Fund carries the same pathogen: 60% of it tracks the FTSE MPF All-World Index, whose US weighting stands at 61.7% (Vanguard, 31 Aug 2026) — same disease, different dosage. Diversification has not vanished, but what you bought is thinner than the label claims.

Second, enforce rebalancing discipline. The 387 fallen stocks are the discount rack. At each annual review, sweep some profits from overweight North American winners back into laggard categories instead of chasing — the +150% vs +50% gap between the Magnificent Seven and the equal-weighted index since 2023 is itself a mean-reversion watchlist.

Third, migrate on fees. You cannot control concentration risk, but you can absolutely control fees. Phase one of full portability is already live, and ECA allows one employer-contribution transfer per calendar year: move North American equity holdings into the market’s cheapest tracker option. A 0.50 percentage-point fee gap, on HK$5,000 a month at 7% gross, is worth HK$568,965 over 30 years in illustrative terms — money that depends on no giant’s share price.

Fourth, respect the unknown-price shield. September’s complete data only arrives in October; hindsight has no trading value. Avoid large switches around the Fed’s 27–28 October meeting, and stage large transfers in tranches so T+1 blind pricing does not eat your rebalancing discipline.

This week’s action list: (1) open eMPF and check what share of your portfolio sits in North American equity funds; (2) measure your total US exposure against the DIS reference portfolio; (3) mark your next ECA transfer date and put your most expensive North American fund on the migration list.

Sources: Barron’s, Dow Jones Market Data, Stocktwits, FactSet (via Barchart), Bloomberg (via Janus Henderson), ATB Wealth, Seeking Alpha, Morningstar, Motley Fool, GUM August 2026 report (via CLS news wire), MPF Ratings 24 Sep 2026. Calculations are illustrative with assumptions stated in the text; past performance is not indicative of future results.