Direct Indexing

Direct indexing is an approach in which an investor holds the underlying securities of an index directly in their own account, in roughly the index's proportions, instead of holding a single fund that tracks it. FINRA describes it as offering "investors a way to purchase many or all the stocks in a specified index, which can include holding hundreds of individual securities." The intended market exposure is much the same as owning the equivalent index fund. What changes is who owns what.

That change of ownership unit is the whole of the idea, and everything claimed for the approach follows from it. A fund shareholder owns one thing: a share of the fund. A direct indexer owns hundreds of things, each acquired at its own price on its own date, each with its own cost basis, and each capable of being sold on its own.

FINRA also notes how the approach became available to ordinary investors: "In the past, this strategy was only available to individuals with over $1 million in liquid assets, often referred to as high net worth investors," and advancements in technology together with the rise of fractional share trading have made it more accessible to retail investors today.

The tax case rests on a single structural fact, and it is easy to state precisely. Inside a fund, the gains and losses on individual holdings are the fund's, not yours. They net against each other within the portfolio and reach you only as a distribution or as a change in the value of your one position. A shareholder in a fund that rose 9% over a year has one holding with one unrealized gain, and no loss to take, even though many of the companies inside the fund fell during that year.

Holding the constituents directly changes what is available. Each position is a separate tax lot in your account, so a holding that has fallen can be sold at a loss whatever the index did, and the proceeds reinvested to maintain the exposure. What the resulting loss is then good for, and the wash sale rule that governs how it is done, is the subject of tax-loss harvesting rather than of this page. The point specific to direct indexing is narrower and is the reason the approach exists: the losses are reachable at all, because the positions are yours rather than the fund's.

Two limits on that argument deserve equal billing. Realizing a loss lowers the basis of what you buy back, so the benefit is largely a deferral rather than an elimination, exactly as it is with any harvesting. And FINRA points out a cost specific to selling individual holdings: "a stock you sell might soon rally back above the cost basis; however, you'll have lost that potential gain as well as altered your portfolio allocation."

The second use is exclusion, and it is more compelling for a narrower group. Because the holdings are individually owned, a company can simply be left out and the remaining weights adjusted. The most defensible reason to do this is not preference but exposure. Someone whose salary, bonus, job security and equity compensation all depend on one employer already has a large economic position in that company, and holding it again inside every broad index fund they own compounds a risk they did not choose. Excluding it is a way of reducing a concentration rather than of expressing a view. Screens applied for other reasons are equally possible and are a matter for the investor.

What the approach costs, and FINRA is direct about both parts. On tracking, it warns that by deviating from the index, "your returns may be different from the index's returns and could potentially be markedly lower." That is the unavoidable consequence of holding something that is not the index: every exclusion, every harvesting sale and every reinvestment introduces a difference, and differences run in both directions. On cost, it warns that "you might also incur higher fees following a direct indexing strategy than you would with a typical passive portfolio," which is worth measuring against the value of the losses being harvested rather than assumed to be smaller than it.

Three practical consequences follow that are easy to underestimate. Holding hundreds of positions means hundreds of tax lots to track, and every corporate action, merger and index change touches the account. The approach does nothing at all inside an IRA or a workplace retirement plan, because losses there are not reportable, so it belongs in a taxable account or nowhere. And the benefit is largest for someone with substantial gains elsewhere to offset and a high rate applying to them, which means its value varies enormously between two investors holding the same portfolio.

You choose an index and a provider, and the account is filled with the index's constituents in approximately their index weights, using fractional shares so that the amounts work at ordinary account sizes. Any companies you want excluded are left out and the remaining weights adjusted. Thereafter the account is monitored: positions trading below their cost basis may be sold to realize losses and replaced so the exposure holds, and the portfolio is periodically brought back toward the index as weights drift and as the index changes.

A hypothetical example of why the unit of ownership matters. Two investors each put $200,000 into the same index at the same time. One buys an index fund; the other holds the constituents directly. Over the year the index returns 9%, so both are up about $18,000 ($200,000 × 0.09).

The fund holder owns one position with an $18,000 unrealized gain, and there is nothing to harvest. Not every company in the index rose, but that is invisible from outside the fund and unavailable in any case.

The direct holder owns hundreds of positions. Suppose, for this illustration, that the ones trading below their purchase price total $14,000 of losses. Selling those realizes a $14,000 capital loss available to offset gains elsewhere, while the proceeds are reinvested so the index exposure continues.

The $14,000 is not free money. The replacement holdings carry a lower cost basis, so a larger gain is waiting when they are eventually sold, and the sales themselves may have moved the portfolio slightly away from the index. What has been gained is timing: tax not paid now stays invested, and it becomes a permanent saving only if the deferred gain is later taxed at a lower rate, donated, or receives a step-up in basis.

Pros

• Losses can be taken on individual holdings even in a year the index rose, which is impossible for a fund shareholder.
• Specific companies can be excluded, which is a genuine way to reduce a concentration for someone already exposed to an employer.
• Fractional shares make it workable at ordinary account sizes, where FINRA notes the approach was formerly confined to investors with over $1 million in liquid assets.
• The positions are owned rather than pooled, so each company's dividends arrive from the holding itself rather than as a single fund distribution.

Cons

• FINRA cautions that returns may differ from the index's and could be markedly lower, which is the unavoidable cost of holding something that is not the index.
• FINRA also cautions that fees may be higher than for a typical passive portfolio, and that cost has to be weighed against the value of the losses harvested.
• The tax benefit is largely deferral, because harvesting lowers the basis of what you buy back.
• A stock sold at a loss can rally back above its cost basis, forfeiting that gain and altering the allocation.
• Hundreds of positions means substantial record-keeping, and the whole approach does nothing inside a tax-advantaged account.