Tax Loss Harvesting - A Loss Is Only Worth What You Spend It On
Joshua Staph
Black vertical line on a white background.
Oct 16 2026 16:41

Tax-loss harvesting is everywhere now. Not long ago it was something a handful of firms did for a handful of wealthy clients, mostly in December and mostly by hand. Today nearly every custodian, robo-advisor and managed account platform offers it, usually automatically, and the statements that come with these accounts have started reporting a number that looks a lot like a return. It goes by different names, like tax alpha, estimated tax savings or losses harvested year to date, but it's presented the same way, as found money.

 

I've used these accounts a fair amount earlier in my career, and two things stuck with me. The first was that the tracking error, meaning how far the account drifted from the index it was supposed to follow, ran higher than what the wholesaler had suggested when the strategy was pitched. The second was that when I dug into the tax alpha figure and tried to look at it on a net basis, it told a much smaller story than the headline did.

 

Neither of those makes tax-loss harvesting a bad idea. It's a real tool, and for some people it's a valuable one. But they made me want to understand what it actually does, as opposed to what it gets credited with doing, and that difference turns out to be most of the conversation.

A quick primer

 

If some of these terms are new, here's what they mean in plain English:

  • Cost basis: what you paid for an investment. It's the starting point for figuring a gain or a loss.
  • Capital gain or loss: the difference between what you sell something for and your cost basis. If you held it more than a year, it's long-term and taxed at lower rates. If you held it a year or less, it's short-term and taxed like your paycheck.
  • Tax-loss harvesting: selling an investment that's down to lock in a loss you can use on your taxes, then buying something similar so you stay invested.
  • Wash-sale rule: if you buy the same or a substantially identical investment within 30 days before or after selling it at a loss, the IRS disallows the loss for now.
  • Loss carryforward: losses you can't use this year roll forward to future years. Beyond offsetting gains, you can use up to $3,000 a year against other income.
  • Direct indexing: owning the individual stocks of an index in your own account instead of a fund, so each one can be harvested separately.
  • Tracking error: how much your account's returns differ from the index it's meant to follow.
  • Step-up in basis: when you die, your heirs' cost basis generally resets to the value on that date, which erases the built-up gain.

What harvesting actually does

 

The mechanics are simple. You own something that's fallen below what you paid for it, so you sell it, which turns a paper loss into a real one you can use on your tax return. Then you immediately buy something similar so the portfolio stays invested. In a direct-indexed account, you own a few hundred of the individual companies behind an index instead of a single fund. Software monitors the portfolio on an ongoing basis, sells positions that have fallen enough to be worth harvesting, and buys similar companies in their place so the account keeps tracking the index.

 

That one sale does two things at once, and nearly all the confusion comes from only counting one of them.

 

The first is the loss itself. I think of it as a coupon. You can spend it against capital gains, and if you don't have enough gains to absorb it, you can use up to $3,000 a year against ordinary income and carry the rest forward.

 

The second is quieter. The replacement you just bought has a lower cost basis than the position you sold, so the gain you avoided paying today hasn't gone anywhere. It's sitting inside the new position, waiting for you. That's an IOU.

 

Most of what gets reported as tax savings is the coupon, valued at a high tax rate, with the IOU nowhere on the page. In a lot of cases, what you've really received is a deferral: you pay later instead of now. Deferral is worth something, the same way an interest-free loan is worth something, but it's a different thing from savings, and the gap between the two can be large.

What your bracket has to do with it

 

One of the first questions I had when I started digging into this was whether you have to be in the top bracket for harvesting to be worth much. The short answer is that your bracket matters a lot, and it matters on both ends of the trade.

 

A harvested loss is worth whatever tax it saves you, and that depends on your marginal rate at the moment you use it. Here's what the same $100,000 loss is worth on a federal return, depending on what it offsets:

What the loss offsets

Federal rate

Tax saved on $100,000

Long-term gains, 0% bracket

0%

$0

Long-term gains, 15% bracket

15%

$15,000

Long-term gains, top bracket

23.8%

$23,800

Short-term gains, top bracket

40.8%

$40,800

 

The top-bracket rates include the 3.8% net investment income tax, which applies above certain income levels. 

 

That's the coupon side. The IOU comes due at whatever rate you're paying when you eventually sell. So the number that really matters is the gap between the rate when you use the loss and the rate when you pay the deferred gain. Your bracket today is only half of it. A high earner who uses losses against short-term gains at 40.8% and pays 23.8% years later keeps the difference for good. Someone in the 15% bracket who uses losses against long-term gains and pays 15% later has mostly just pushed the bill back. And someone in the 0% bracket usually shouldn't be harvesting losses at all. They're often better off doing the opposite, realizing gains while they're tax-free to raise their basis.

 

So no, you don't strictly need to be in the top bracket. But you do need either a real gap between today's rate and tomorrow's, or a way to make the IOU disappear entirely. Without one of those, the benefit shrinks toward the time value of money.

 

Vanguard's research on this, by Paradise, Khang and Dickson, found that current and future tax rates were the biggest single factor in how much harvesting turned into actual tax savings. It recommends against harvesting when your gains fall in the 0% bracket, and warns that if rates are expected to rise, harvesting may only subtract value. Keep in mind that Vanguard offers direct indexing itself. This isn't a skeptic's argument.

 

The same paper shows how much the headline depends on assumptions. Its median estimates of the long-term benefit ranged from about 0.47% to 1.27% a year when everything went right: daily monitoring, a few hundred individual stocks, regular new contributions, and every dollar of tax savings reinvested back into the portfolio. With quarterly checks, a single fund, no new money, and half the savings reinvested, the benefit fell to roughly 0.04% to 0.13% a year. Most people I know don't reinvest their tax savings. They spend them, which is perfectly reasonable, but it's not what the bigger number assumes.

 

When the coupon turns into real money

 

There are a handful of situations where a harvested loss becomes a permanent benefit instead of a postponed bill:

  • You spend it at a higher rate than you'll eventually pay. Offsetting short-term gains or ordinary income today, which are taxed like your paycheck, and paying long-term capital gains rates on the IOU years from now.
  • The IOU never comes due. The position is held until death and your heirs receive a step-up in basis.
  • You give the appreciated shares away. Donating them to charity or a donor-advised fund instead of selling them.
  • You realize the deferred gain in a low year. For example, a year when you fall into the 0% capital gains bracket.
  • The losses absorb a gain you were going to take anyway. A business sale, a property sale, or unwinding a concentrated stock position.

Outside those situations, the benefit is mostly the time value of the deferral. Sometimes it's less than that, if the deferred gain eventually gets realized at a higher rate than the one you saved, whether because your income is higher later or because the tax code changes.

 

The best-known academic study on this, by Chaudhuri, Burnham and Lo, found harvesting may have added about 1.08% a year before costs from 1926 to 2018, falling to about 0.82% once the wash-sale rule was respected. That's a real number. But it assumed losses were used against short-term gains taxed at 35%, which is exactly the first situation on the list above. The value leans heavily on having the right kind of gains to spend the coupons on.

 

One local note for my Pennsylvania readers: as I understand the state rules, Pennsylvania doesn't let capital losses carry forward to future years, and it doesn't let them offset other kinds of income. So whatever a harvested loss is worth on your federal return, it's likely worth less on your state return than you might assume.

 

The part about dying

 

This is where it gets counterintuitive, because two rules seem to pull in opposite directions.

 

On one side, the step-up in basis at death wipes out built-up gains. If you hold a position until you die, the IOU you created by harvesting is simply forgiven. On the other side, unused capital losses don't survive you. A loss carryforward can be used on your final tax return, and then it's gone. For married couples, the surviving spouse generally can't use the deceased spouse's share of the carryforward after the year of death (The Tax Adviser has a good walkthrough of how that works).

 

So is death good for harvesting or bad for it? I went back and forth on this until I realized the two rules are acting on different things. The step-up cancels the IOU. Death expires the coupons. If you're going to hold until death, the IOU side of harvesting never costs you anything, and the only question left is whether you'll actually spend the coupons while you're alive.

There's one more piece that settles it. If you never harvest at all, an unrealized loss you're holding at death doesn't get a step-up. It gets a step-down, and it disappears anyway. So harvesting isn't what creates the problem of unused losses. It just gives you a chance to use losses that would otherwise vanish.

 

The real failure is ending up with a large pile of carryforwards and nothing to spend them on. At $3,000 a year against ordinary income, a $400,000 carryforward would take more than a century to use up. A big carryforward late in life is unspent inventory.

 

The harvest doesn't last

 

This one is mostly common sense. When you first fund an account, every position starts at its purchase price, and plenty of them will dip below it. A few years later, after the market has done what it usually does over time, most positions are sitting well above their cost, and there's very little left to sell at a loss.

 

The research backs this up. A 2023 paper by Liberman, Krasner, Sosner and Freitas found that in a direct-indexed portfolio, harvested losses added up to about 13% of the original investment in year one and passed 20% by year three, then stayed roughly flat from years six through ten. The estimated tax benefit averaged about 0.69% a year over the first five years and about 0.09% a year over the next five.

 

An earlier paper by Sosner, Gromis and Krasner goes a step further. With no new money going in and only long-term gains to offset, the benefit fell to roughly zero after about five years. With ongoing charitable giving of appreciated shares, it held up for much longer. That's the coupon idea showing up in the data: the harvest stays valuable when there's somewhere to spend it. The authors' figures also subtract an estimate of the future tax on the gains building up underneath, which is exactly the net view I was trying to get to years ago. If your statement shows a tax savings number, it's fair to ask whether it does the same.

 

Both papers come from AQR, and you should know that AQR sells the strategy that's meant to pick up where direct indexing leaves off, which I'll get to in a later piece (Tax Aware Long/Short). That doesn't make the finding wrong. It's mostly arithmetic, and it lines up with what I saw in practice. But in this business, it's always worth asking who did the research and what do they sell.

 

The other cost is drift. A direct-indexed account doesn't hold the index exactly. It holds a few hundred of its companies, swaps in substitutes every time it harvests, and may exclude stocks you've asked it to avoid. Each of those decisions moves the account a little further from the benchmark, and the more aggressively it harvests, the further it can wander. Drift can work in your favor or against you, but it's rarely zero.  I personally haven't really seen it work in your favor either.  

 

The math can be a little bit of a problem, and the behavior can be more. In a year when the index is up 25% and your account is up 22%, the tax savings on your statement can feel pretty abstract next to the return you didn't get. That gap is hard to sit through, and anyone who gives up on the strategy because of it pays the deferred taxes all at once on the way out. I probably could have boldened this whole sentence.  It probably stings the most.

 

What happens after the harvest fades is the part nobody puts in the pitch. The management fee and the tracking error are still there, and you now own several hundred individual positions, most of them with built-up gains, so leaving for something simpler means paying the tax you spent years deferring. 

 

How much you end up harvesting also depends on what the market happens to do in your first few years. Invest right before a rough stretch and there's plenty to harvest. Invest right after one and there's much less. Nobody controls that, so any projection of tax savings is partly a guess about luck.

 

Where it actually pays off

 

None of this means harvesting is a gimmick. Like I said above, I've utilized these strategies for clients. For the right person, it can be one of the more useful tools available, and many of the people I work with fit at least one of these descriptions.

 

The clearest case is someone with large gains coming. If you're likely to sell a business, sell a property outside a 1031 exchange, or work your way out of a concentrated stock position, harvested losses have an obvious place to go. In that situation the coupons get spent, and the benefit is real and fairly easy to measure.

 

Charitable giving is another. If you give regularly, you can give away the positions with the largest gains and keep harvesting losses from the rest, which keeps the account productive long after a plain harvesting strategy would have stalled.

 

The same goes for people who expect to hold their taxable investments for life, since the step-up erases the IOU. And it helps a lot if you keep adding to the account. Every new deposit buys positions at today's prices, which gives the harvest fresh material to work with, the same way the first few years did. Vanguard's best-case numbers assumed regular contributions for exactly that reason. A large one-time deposit, like an inheritance or the proceeds from a liquidity event, gets the same early-years boost.

 

For executives with equity compensation, there's a benefit that has nothing to do with taxes. A direct-indexed account can exclude your own company's stock, and even its sector, so the rest of your portfolio isn't quietly doubling down on the exposure you already have through your job.

 

And there's a behavioral benefit I don't want to dismiss. In a sharp decline, harvesting gives you something productive to do with the fear, and that can make it easier to stay invested when it matters most.

 

For people with very large gains to absorb, there are more aggressive strategies that use leverage and short positions to keep generating losses long after a direct index runs dry. Those come with real costs and risks, and they deserve their own piece, which is coming.

 

Questions worth asking

 

Whether you're considering one of these accounts or already have one, these are the questions I'd want answered:

  • What am I going to spend these losses on, and roughly when? If there's no good answer, the coupons may never be worth much.
  • Is the tax savings number on my statement net of the gain I'm deferring? And what tax rate is it assuming?
  • What does year six or seven look like? When there's little left to harvest, what am I paying for, and what would it cost to leave?
  • How far has this account actually drifted from its index, and would I sit through that drift during a bad stretch?
  • Am I accidentally washing my own losses? That includes purchases in a spouse's account, a dividend reinvestment, an employee stock purchase plan, or shares of your own company's stock vesting from RSUs. If the repurchase happens inside an IRA, the loss isn't just delayed. It's generally lost for good.

Spending the coupon

 

Tax-loss harvesting is a real tool with a real job. Its job is to produce losses you can use, and a loss is only worth what you eventually spend it on. When there's a business sale on the horizon, a concentrated position to unwind, or a habit of giving, it can be very much worth having. When there isn't, a lot of what shows up on the statement as savings is a bill you've moved to later, wrapped in an account that gets harder to leave every year.

 

I care more about you making a good decision than which decision you make. If you own one of these accounts, or someone is pitching you one, the question to start with is simple: what are we going to spend these losses on? If the answer is clear, you're probably in the right place. If nobody can answer it, that's worth a conversation.

 

Joshua Staph, CIMA®


Founder, Verak Private Wealth · joshua@verakprivatewealth.com · verakprivatewealth.com

The views expressed are those of the author and are for informational and educational purposes only. Not investment, tax, or legal advice. Tax treatment depends on your individual circumstances and may change under future legislation; consult a qualified tax advisor before acting. Any figures are hypothetical and not a projection or guarantee of future results.

 

Securities offered through Cambridge Investment Research, Inc., Member FINRA/SIPC. Advisory services through Cambridge Investment Research Advisors, Inc., a Registered Investment Adviser. Verak Private Wealth LLC and Cambridge are not affiliated.

 

This article is for educational purposes only and is not tax, legal or investment advice. Tax rules are complex and change over time, and your situation may differ, so please consult your tax advisor before acting. References to research and to firms, including AQR and J.P. Morgan, are for informational purposes and are not a recommendation of any firm, product or strategy. Tax-loss harvesting and direct indexing involve risks, including tracking error, and may not be suitable for all investors. Past performance does not guarantee future results.

Sources