Direct indexing means you own the individual stocks inside an index instead of owning a fund that tracks it. You buy 300 of the 500 S&P names directly in a brokerage account, weight them to mimic the index, and hold them yourself. The entire point is tax control: when a stock drops, you sell it for a loss, buy something similar, and use that loss to offset gains elsewhere. An index fund cannot do this for you. A direct index portfolio can, constantly, in the background, without you doing anything.
An S&P 500 ETF holds 500 stocks inside one wrapper. You own one share, the fund owns the rest. You cannot sell just the losing positions inside it, because you do not own them individually. Direct indexing strips the wrapper away. A custom portfolio is built to track the index using most or all of its constituents, held directly in your name, inside your own account.
The tracking is not perfect and it does not need to be. Software rebalances the portfolio periodically to stay close to the benchmark's return and sector weights. Most platforms get within a few basis points of the index over a full year. You give up a sliver of precision. In exchange you get something an ETF structurally cannot offer: a portfolio made of hundreds of separate tax lots, each one a lever you can pull.
This is the actual reason direct indexing exists. Direct indexing tax loss harvesting works because markets are volatile at the individual stock level even when the index itself is flat or up for the year. In any given year, dozens of S&P 500 names will be down 10%, 20%, sometimes 40%, even while the index posts a positive return. Each one of those declines is a harvestable loss if you own the stock directly.
The mechanic is simple. You sell the stock that's down, you realize a capital loss, and you immediately buy a correlated but not identical replacement to keep your market exposure roughly the same. Thirty-one days later, once the wash sale window closes, you can buy back the original name if you want. The realized loss goes on your tax return and offsets capital gains, or up to $3,000 of ordinary income per year if you have no gains to offset. Unused losses carry forward indefinitely.
Do this across a diversified basket of stocks, every quarter or every month, and the harvested losses compound. A well-run direct index account can generate tax losses worth 1% to 3% of the portfolio's value per year, especially in the first two years when volatility hasn't been picked over yet. That's not a return. It's a deferral of tax liability, which functions like a return if you'd otherwise be paying tax now on gains elsewhere.
An ETF is cheaper to run and easier to understand. Expense ratios on S&P 500 ETFs sit near 0.03%. Direct indexing platforms charge 0.20% to 0.40% annually for the management and tax software. That fee gap is real and it matters if you're not going to use the tax-loss harvesting feature at all.
| Feature | ETF/Index Fund | Direct Indexing |
|---|---|---|
| Annual cost | 0.03% to 0.10% | 0.20% to 0.40% |
| Tax-loss harvesting at stock level | Not possible | Built in |
| Custom exclusions (single stock, sector) | Not possible | Yes |
| Minimum account size | None | Usually $100,000+ |
| Tracking precision to index | Near exact | Close, not exact |
The exclusion feature deserves more attention than it gets. If you work at a tech company and already hold a concentrated stock position, a direct index account can be built to exclude that one company entirely, so your diversified holdings don't double down on the same risk sitting in your equity comp. An ETF can't be edited that way. You take the whole basket or none of it.
Direct indexing earns its fee for people with large taxable gains to offset: someone who sold a business, someone sitting on concentrated stock, someone with a high W-2 income and a taxable brokerage account they're actively funding. The harvested losses have somewhere useful to go. Without gains or high ordinary income to offset, harvested losses just pile up as an unused carryforward, and you paid extra basis points for nothing.
It also barely helps inside a 401(k) or an IRA. Tax-loss harvesting only matters in a taxable account, because retirement accounts already grow tax-deferred or tax-free. Anyone pushing direct indexing for your Roth IRA is selling you a feature you can't use. That's a real thing that happens, and it should make you suspicious of the advisor, not the strategy.
Account size is the other filter. Below roughly $100,000, the fixed operational overhead of running hundreds of individual positions eats into any tax benefit, and the tracking error becomes more noticeable relative to the account's size. Under that threshold, a plain ETF and a disciplined once-a-year harvest of a few positions gets you 80% of the benefit for a fraction of the cost.
The fee schedule is the visible cost. The invisible one is complexity at tax time and at exit. Direct index accounts generate hundreds of transactions a year. Your tax preparer will bill you more for the extra work, or you'll spend hours reconciling a 1099-B with more line items than a small business's ledger.
Exit cost is the bigger issue. Because the account holds individual stocks with staggered cost bases, some purchased high and some low, unwinding the whole position at once can trigger a large tax bill if a chunk of the holdings have appreciated since purchase. Direct indexing is built to be held for years, ideally decades, so the harvested losses keep compounding and the eventual step-up in basis at death (for US taxpayers) can erase the deferred gains entirely. Treat it as a short-term parking spot and you'll likely come out behind an ETF once you add up the fees and the tax mess of unwinding it.
Wash sale rules also get harder to manage at scale. Selling a stock for a loss and buying a "similar but not identical" replacement sounds simple in theory. Do it across 300 positions a year, across multiple accounts including a spouse's IRA, and it becomes very easy to accidentally trigger a wash sale without realizing it. Good platforms track this automatically. Cheap or DIY versions of direct indexing often don't, and the IRS doesn't care that you didn't mean to.
Match the strategy to a real need before you sign up. If you have a six or seven figure taxable account, expect large capital gains in the next few years (a business sale, RSU vesting, an inheritance of appreciated stock), and you're in the top two federal tax brackets, direct indexing is worth the added fee. The harvested losses will have real gains to offset, and the fee is small next to the tax saved.
If your investable assets are mostly in a 401(k) and a starter taxable account under $50,000, skip it. Buy the ETF, hold it, and harvest losses manually once a year during a downturn if you want the same effect at near-zero cost. The strategy scales with wealth and tax liability, not with how sophisticated it sounds. Most people asking about direct indexing are curious about the mechanism, not actually positioned to benefit from it yet, and that's fine. It's a tool for a specific balance sheet, not a universal upgrade.
Usually not. Below roughly $100,000, the added fee of 0.20% to 0.40% a year outweighs the tax benefit, and the fixed cost of managing hundreds of small positions makes tracking error worse. A standard low-cost ETF with a manual once-a-year loss harvest gets most of the same benefit for far less money.
No, and it is not designed to. The portfolio tracks the index closely rather than beating it. The value comes from the tax losses generated along the way, not from outperforming the benchmark itself.
You can, but the main benefit does not apply. Tax-loss harvesting only matters in a taxable brokerage account, since retirement accounts already grow tax-deferred or tax-free. Paying an extra fee for direct indexing inside a Roth IRA gets you a feature you cannot actually use.
An index fund holds all the stocks inside one shared wrapper, so you own shares of the fund, not the individual companies. Direct indexing holds the same stocks separately in your own account, which lets you sell individual losing positions for a tax deduction while keeping the rest of the portfolio intact.
The biggest risk is triggering an unexpected wash sale across dozens or hundreds of trades, which disallows the loss deduction you were trying to generate. A close second is unwinding the whole account at once later on, since staggered cost bases can create a large, unplanned tax bill if not managed carefully.