Long-only direct indexing vs. tax-aware long-short direct indexing

Let’s get on the same page about some definitions (if this is not new, feel free to skip ahead)…

A quick definition of direct indexing

You own the index name by name. Instead of one fund share, you hold a few hundred individual stocks weighted to track the benchmark. Your market exposure looks the same as a fund. Your tax options do not.

Why would someone use a direct indexing strategy?

When one holding dips below what you paid, the account sells it, banks the capital loss, and replaces it with a similar stock so the portfolio keeps tracking the index. A fund can’t do that. It has one price for the whole basket, so a loss in any single name stays buried inside it.

Those banked losses offset capital gains elsewhere on your return: the embedded gain in an appreciated holding you trim, a business sale, the unwind of a concentrated position.

A harvested loss is only ever worth what you can put it against, though. Match it to a gain and your tax bill shrinks; the size of that benefit scales with your bracket.

There’s a small backstop even when you have no gains to match. Up to $3,000 of net harvested losses can offset ordinary income each year under federal rules, and anything past that carries forward indefinitely, waiting for a gain to absorb it.

So, if you produce gains, a machine that produces losses is useful.

Side note: Harvesting is the headline, but it isn’t the only reason to own the index name by name. Holding the names individually lets you leave some out. If a big share of your net worth already rides on one company–say a pile of employer stock like SpaceX shares–you can track the index with that name carved out, so you aren’t doubling the exposure your paycheck already gives you. The portfolio still tracks the benchmark; it just stops piling onto a risk you already carry. The same logic covers a whole sector you’re heavy in, not only a single name. And when the concentration is in something the index doesn’t hold, you can trim the public names most correlated to it instead.

What’s the catch with direct indexing?

Long-only direct indexing (that’s what we’ve been talking about) harvests losses well in its early years, then slows down once most of your holdings sit above what you paid for them.

In other words, harvesting capacity is front-loaded. Initially, every lot sits right at its purchase price, so any dip in any name creates a loss to take. In a rising market, more and more lots drift above what you paid, and a stock now has to fall below your cost before there’s anything to harvest. By year four or five of an up market, an account that once threw off big losses may produce very little. This is called “ossification” (in the tax-aware investing circles… probably not a word you’d use with your friends).

Fresh contributions buy you more runway

One simple way to get around “ossification”: keep adding money.

Every new dollar buys lots at today’s prices, and new lots are the ones with a chance of dipping below their cost. A household steadily moving RSU proceeds, bonuses, or monthly savings into the account keeps restocking the shelf the harvester picks from.

Contributions don’t fully solve the decay–older lots still age out–but for a household still accumulating, they can extend long-only’s productive years by a lot. If your contribution stream is large relative to the size of the account, that can keep the loss harvesting going for some time.

How does long/short direct indexing differ from long-only?

Long/short direct indexing pairs the long index-like portfolio with a short book: positions that profit when specific stocks fall.

A common shape is 130/30, meaning $130 of long exposure and $30 of short exposure for every $100 invested.

Notice the long side runs to $130, above the $100 you’d hold in a long-only account. That extra $30 is the levered part. In simplified terms, the short book helps finance it: selling $30 of stock short generates cash that can be used as part of the collateral and financing structure for the additional $30 on the long side.

Net market exposure still comes back to roughly $100 — $130 long minus $30 short — so the account can be designed to maintain equity exposure similar to a long-only portfolio.

But it is not the same as owning a plain index fund, and it is not automatically as index-like as long-only direct indexing. What rises is the gross: $160 of stock exposure standing on $100 of your capital, spread across a larger long book and a new short one. That brings more moving parts: financing costs, short-borrow costs, margin rules, collateral requirements, tracking error, and manager execution risk.

The larger long book adds something of its own: more individual names that can dip below cost, so more to harvest on the long side than a $100 long-only book would hold.

The short book adds another surface to harvest from. A short position loses money when the shorted stock rises, and that loss may be harvestable when the short is closed. So the strategy does not have to wait only for the broad market to fall. In a rising market, some stocks still fall, and some short positions move against the portfolio, which can create harvestable losses.

The long/short structure creates more persistent tax-loss harvesting than long-only.

But “more persistent” does not mean guaranteed. The result depends on market dispersion, manager decisions, wash-sale management, financing costs, fees, tax rates, and whether you actually have gains to offset. Some of the tax value also comes from managing when gains are realized, not just manufacturing losses.

Managers and researchers who model these strategies show the long/short version generating substantially more harvestable losses than long-only over the same span, especially in the later years when long-only has tapered off. Read those figures as modeled, not as anyone’s actual account, because the assumptions behind a model can flatter the result.

Some managers push the exposure further, to 150/50, 200/100, or even 250/150. More short exposure can mean more harvestable losses, but it also means more gross exposure, more financing cost, more tracking error, more manager dependence, and more risk that the pre-tax investment result does not justify the structure.

The clearest case for levering up

Is a large, known gain coming down the road — an IPO unlock, a business sale, a sizable secondary, or a planned multi-year sell-down of a concentrated position. When you can see a seven- or eight-figure gain ahead of time, you can run a long/short strategy in front of it and deliberately build a stockpile of losses to set against it.

Depending on the size of the gain and your bracket, the tax saved can run into six figures.

What makes or breaks that plan is lead time. Harvested losses accumulate, so the longer the account runs before the gain lands, the more time the manager has to fill the tank. Start three years out and the harvester has three years to work; start three months out and it has…a lot less than 3 years.

The calendar can swing this more than people expect: a gain that lands on January 1 instead of December 31 hands you essentially a full extra tax year to prepare. That is the example pushed to its limit, since closing dates rarely turn on a single day, but the direction holds — when the gain falls decides how much runway you had to get ready for it.

There is no free lunch.

The short book has to be financed, gross exposure rises, and fees run higher than long-only direct indexing, which already costs more than a plain index fund. Short positions can move against you. Borrowing securities is not free. Margin and collateral requirements matter. You’ve also tied yourself to one platform and one manager’s execution.

The exit is harder, too. A short book doesn’t usually transfer in kind the way a basket of long stocks can, so leaving can mean unwinding it. And unwinding it can mean recognizing the very gains you spent years deferring.

There are also tax details that make this more complicated. Short-sale gains and losses generally show up when the short is closed. Short-sale losses can run into wash-sale issues. Payments in lieu of dividends and short-sale expenses can have their own tax treatment. And if you are shorting around a concentrated appreciated position, constructive-sale rules can matter.

AQR, which manages these strategies and publishes research on them, makes a point that cuts both ways: the pre-tax performance of the long/short positions largely determines whether the whole thing succeeds. If the strategy adds value before tax, the tax benefit stacks on top. If it subtracts value before tax, the tax benefit may only offset a loss you didn’t need to take in the first place.

For an investor without large ongoing gains, fees and financing costs can eat up much of the tax savings, and eventually liquidating the portfolio can claw back a good chunk of whatever lifetime benefit was left.

So is direct indexing worth it?

For a household with recurring gains and a taxable account of meaningful size, long-only has a credible case. The long/short version fits a narrower set: large upcoming gains and comfort with the structure. Plenty of households are well served by neither, and an index fund paired with a high savings rate is nothing to apologize for.

How I think about it

Start from the gains your household will actually produce over the next five to ten years. Model the benefit conservatively, with any hypothetical numbers labeled as such, and weigh it against the fees, the financing cost, the platform dependence, and the exit.

The answer can be a long/short structure, long-only, or a reminder that your current setup already does the job. If you’re staring down a multi-year sell-down and wondering which bucket you’re in, that’s a conversation worth having: stoneholtwealth.com/get-started.