If you hold one S&P 500 ETF and the index drops 3%, you have exactly one tax-loss harvest opportunity: sell the ETF, buy a "substantially different" one (VOO ↔ IVV is the classic dance), and move on.
If you hold the underlying 60–80 stocks that make up the same index, the picture is dramatically different. Even on a day the index is flat, half your positions are probably down 1–8% on their own. Each one is an independent harvest opportunity.
This article quantifies the gap. It's the single biggest reason direct indexing exists.
The "harvest surface" idea
Imagine your portfolio as a row of light bulbs. Each bulb is a position. A bulb is "lit" (harvestable) when its current price is materially below your cost basis on a recent enough lot.
- Hold one ETF? You have one bulb. It's lit only when the entire index is down — a few times per year.
- Hold 60 stocks tracking the same index? You have 60 bulbs. On any given day, some fraction of them is lit independently of what the index did.
That fraction is what we'll model. The simulator below lets you play with three things — index volatility, index drift, and the dispersion across stocks — and watch the harvest surface light up.
Watch a year unfold
The chart below simulates 252 trading days for two portfolios holding the same dollar amount: one S&P 500 ETF (top row) and 60 individual S&P 500 stocks (bottom grid). Each cell is one day; cells go red when that position becomes harvestable on that day. Press Run year to animate.
Simulation: $1M portfolio, GBM for the index plus per-stock idiosyncratic noise, 30-day wash-sale lockout after each harvest. Defaults: 18% vol / 20% dispersion / 5% loss trigger reproduce realistic 2022-style behavior.
Why the gap is so big
The mechanic that drives this is idiosyncratic volatility — the part of a single stock's daily move that isn't explained by the broad market. Even when the index is up 1%, half the names in it might be down. Below the index, AAPL went sideways while NVDA dropped 8% and XOM rose 3%. Each of those divergences is a potential harvest.
An ETF averages all those divergences away by construction. You bought the average. There's nothing to sell separately.
What you give up
Holding 60 stocks isn't free. Three real costs:
- Tracking error. A 60-stock S&P 500 sample doesn't perfectly match the index. Well-designed direct portfolios hold tracking error to about 0.5–1.5% per year. Across 30 years, the expected drag is small relative to the tax savings — but it isn't zero.
- Cognitive load. Fifteen positions feel manageable. Sixty look like a job. Software is the answer here, but it has to actually be good. Approving 30 trades a quarter by hand is a different experience than approving one.
- Coordination across accounts. Wash-sale rules span every account you control (and your spouse's). Five separate brokerages with overlapping holdings is harder than one.
When the ETF actually wins
Direct indexing isn't magic. It loses to a single ETF when:
- Your account is below ~$50K. Per-trade commissions used to dominate here; even on $0-commission brokers the per-name minimum size makes 60 holdings impractical (you'd own one share each).
- Your tax bracket is low. If you're paying 0% on long-term gains (the bottom federal LTCG bracket extends to roughly $94K MFJ in 2026), the harvest doesn't save you anything because you weren't going to pay tax anyway.
- Your account is in a tax-deferred wrapper (IRA, 401(k), HSA). Realized losses inside a wrapper do nothing. TLH is a taxable-account game.
The TL;DR
If you have a taxable account above $100K, are paying meaningful capital-gains tax, and don't mind looking at 60 ticker symbols instead of one — direct indexing harvests roughly 5× more loss than a single ETF, every single year. The expected tax-alpha at typical high-bracket rates is 0.5–2.0% per year. Over 30 years, on a seven-figure account, that compounds into the high six figures.
It's the most sustainable edge available to a self-directed investor that doesn't require predicting a single thing about the market.