Tax Losses Can Be Valuable. Losing Money Isn't.

“I’m down forty grand on it, but at least it’s a write-off.” You hear some version of this about a speculative stock, a friend’s startup, a crypto position, or a side business that never turned a profit. The sentence is usually offered as consolation, and sometimes as a reason to keep going.

The tax part of that sentence is often true. The consolation is mostly arithmetic that nobody finished. A deduction returns a fraction of what you lost, sometimes a small fraction, sometimes decades later, and sometimes nothing at all.

We covered the general mechanics of deductions, and rental depreciation in particular, in what a rental write-off is actually worth. This guide is about investment losses: what the tax code gives back when a position goes bad, what it withholds, and why the write-off tends to feel larger than it is.

Compare the loss with not losing

The IRS’s own definition does most of the work. “A deduction is an amount you subtract from your income when you file so you don’t pay tax on it,” while “a credit is an amount you subtract from the tax you owe.”¹ A fully usable deduction of L at marginal rate t is worth t × L. The after-tax loss is L × (1 − t). With any rate below 100%, the loss stays a loss.

When someone says a $20,000 loss “only cost” $15,000 after tax, the statement can be accurate. The error is in the comparison. The tax benefit makes the loss smaller than a loss with no deduction. The choice an investor faces is usually between losing $15,000 after tax and losing nothing, and on that comparison the deduction does not change which option is better. Tax treatment changes the size of an outcome. For a loss that the tax code recognizes one for one, it cannot change the sign.

The objective that follows is to maximize expected after-tax wealth. A person who pays $500,000 of tax on a $2 million gain ends up far richer than one who pays nothing because the investment went to zero. Taxes belong inside the calculation. They make a poor target for it.

What a $100,000 capital loss is worth with no gains to offset

The t × L shortcut also assumes the deduction is usable in full, right away, at your ordinary rate. For an individual’s capital loss, that is often false.

Capital losses first offset capital gains without limit. Whatever is left can reduce ordinary income by at most $3,000 a year ($1,500 if married filing separately), and the rest carries forward “until it is completely used up.”² The $3,000 figure is written into the statute with no inflation adjustment,³ and the IRS’s 2026 inflation adjustments leave it unchanged.⁴ The carryforward also has an end date that people forget: a decedent’s unused capital loss can be deducted only on the final return, and “the decedent’s estate cannot deduct any of the loss or carry it over to following years.”²

Here is what that does to one modeled investor. They realize a $100,000 net capital loss this year and never have a capital gain to absorb it, so it comes off ordinary income $3,000 at a time: 34 tax years, starting this one. We discount each year’s tax saving at 5% and use federal rates only.

Assumptions: $100,000 net capital loss, no capital gains in any later year, $3,000 deducted against ordinary income in each of years 0 through 32 and $1,000 in year 33, constant marginal rate, 5% nominal discount rate, no state tax. The right-hand column stops after 20 deductions ($60,000 used) and forfeits the remaining $40,000 at death.

Under these assumptions the loss is worth about half its face value in tax, at every bracket, because the drag comes from the $3,000 pace and the discount rate. Even the face value is a minority of the loss: the 37% investor is out $100,000 and gets back $18,725 in today’s dollars.

Gains let you use the loss faster, often at a lower rate. If the same investor had $100,000 of long-term gains to absorb the loss this year, the loss would be worth up to $23,800 at once (20% plus the 3.8% net investment income tax), and less for anyone in the 15% capital gains bracket. The loss offsets gains at capital gains rates, which for most investors are lower than their ordinary rate. Short-term gains are the exception, taxed as ordinary income, and a loss that offsets them is worth the full marginal rate. None of these cases gets near $100,000.

Losses the tax code ignores

Some losses produce no tax benefit at all, which makes “at least it’s a write-off” simply false.

• Personal-use property. “Losses from the sale of personal-use property, such as your home or car, aren’t tax deductible.”⁵ An individual can deduct losses from a trade or business, from a transaction entered into for profit, and from casualty and theft, and nothing else.⁶ The boat, the collectible car you drove, and the house you lived in do not qualify.
• Hobbies. An activity not engaged in for profit gets no deductions beyond the narrow allowance in section 183,⁷ and the suspension of miscellaneous itemized deductions, which is where hobby expenses used to land, was made permanent in 2025.⁸ A side business that the IRS treats as a hobby produces taxable income when it earns money and no deduction when it loses money.
• A wash sale into your IRA. Sell a fund at a loss in a taxable account, buy it back inside your IRA within 30 days, and the loss is disallowed with no increase to the IRA’s basis.⁹ An ordinary wash sale defers the loss. This one destroys it.

Two rules run the other way and are worth knowing. Stock in a qualifying small business, bought directly from the company, can generate an ordinary loss of up to $50,000 a year, or $100,000 on a joint return, instead of a capital loss.² That is the section 1244 treatment, and it is the closest the code comes to making a speculative loss fully usable. And a security that becomes completely worthless is treated as sold on the last day of the tax year, so you do not need a buyer to claim the loss.² Neither rule makes the loss a good outcome. Both make the tax benefit closer to its face value.

The one era when buying losses paid

There was a period when individuals bought losses on purpose, in bulk, and it made sense for them. It is the clearest evidence of what it takes for a tax loss to be worth more than it costs.

Before the Tax Reform Act of 1986, the top marginal rate on ordinary income was 50%, 60% of a long-term capital gain was excluded from tax, and losses from limited partnerships in real estate, oil and gas, and equipment leasing could offset wages and other income. A top-bracket investor could deduct accelerated depreciation and interest at 50 cents on the dollar. In Samwick’s worked real estate example, the gain at sale was taxed at an effective 20%: 40% of the gain, at 50%.¹⁰ The tax loss was larger than any economic loss, and in many deals there was no economic loss: Andrew Samwick notes that these investments “can create such tax losses even while generating positive economic income and cash flows.”¹⁰

In IRS Statistics of Income tabulations, limited partnerships with losses reported $23.5 billion of losses in 1981 and $52.3 billion in 1986, yet the number of loss partnerships and the number of partners in them kept rising. Susan Nelson of Treasury and Tom Petska of the IRS called this “counter to conventional economic motives which would have predicted resources (firms and investors) expanding in profitable activities and declining where losses were incurred. The observed patterns are instead consistent with tax sheltering motives.”¹¹ By one IRS tabulation, among 1986 taxpayers with at least $250,000 of positive income who paid an average tax rate of 5% or less, partnership losses offset over 40% of that income.¹⁰

The 1986 act cut the top rate to 28%, repealed the capital gains exclusion, and added the passive activity loss rules. Public limited-partnership sales fell from $13.1 billion in 1986 to $2.6 billion in 1992.¹⁰ The passive loss rules usually get the credit, but Samwick’s own estimates find them of secondary importance next to the rate cut and the loss of the capital gains exclusion.¹⁰ Either way, the buyers were responding to a gap between the tax loss and the economic loss, and that gap is what closed.

An ordinary investor selling a losing stock today has no such gap. The tax loss equals the economic loss, the benefit is at most the marginal rate, and the $3,000 limit often shrinks it further. The shelter era shows the conditions under which a loss can be worth buying, and a brokerage account in 2026 meets none of them.

Why a write-off feels like a refund

Part of the problem is plain confusion about how deductions work. In a 2024 national poll for the Tax Foundation, respondents were asked whether a $1,000 deduction or a $1,000 credit is worth more to someone taxed at 10% on $10,000 of income. Only 47% of respondents with postgraduate education answered correctly, and 24% of those with a high school education.¹² Someone who half-believes a deduction is a refund will overvalue every write-off they are offered.

The rest fits three findings from behavioral finance. We know of no study that directly measures people using deductions to rationalize bad investments, so treat what follows as consistent with the evidence rather than proven by it.

• Mental accounting. Richard Thaler describes how people track money in separate accounts and evaluate each against its own reference point, which “violates the economic principle of fungibility.”¹³ A write-off lets a failed account close with something in the plus column, even though the household balance sheet is smaller.
• Sunk cost. Hal Arkes and Catherine Blumer randomly assigned theater season-ticket buyers to full price or a discount, and the full-price group attended more plays.¹⁴ Money already spent pulls on the next decision. “I’ll keep it for the write-off” can be that pull with a tax justification attached.
• The disposition effect. Terrance Odean’s study of 10,000 brokerage accounts found investors realized gains about 1.5 times as readily as losses, and that the winners they sold beat the losers they held by 3.4 percentage points over the following year. December was the exception, when tax-motivated selling took over.¹⁵ We cover this at length in yes, an unrealized loss is still a loss.

Where tax losses add value

Using losses well comes down to sequence. Tax-loss harvesting is legitimate because the economic loss comes first and the tax decision second.

Suppose a broad index fund you bought for $100,000 falls to $80,000. The $20,000 is gone whether you sell or not. You now face two separate questions: do you still want this market exposure, and can you rearrange the position so the decline becomes useful at tax time? Sell, buy a similar fund that is not substantially identical, and you hold a tax asset while keeping the exposure you wanted. George Constantinides showed in 1984 that a tax system based on realization gives investors a timing option, to realize losses and defer gains, and that the option has real value.¹⁶ Investors respond to it: comparing the same people’s taxable and tax-deferred accounts, Ivković, Poterba, and Weisbenner found tax-loss selling throughout the year, strongest in December.¹⁷

The value is real and bounded. A historical simulation over the 500 largest U.S. stocks from 1926 to 2018, at assumed rates of 35% short-term and 15% long-term, found 1.08% a year of tax alpha before transaction costs, falling to 0.82% with the wash-sale rule enforced.¹⁸ A Vanguard study found that the investor’s own profile, meaning tax rates and whether gains exist to absorb the losses, drove roughly 60% of the variation in outcomes.¹⁹ Harvesting also lowers your cost basis by the amount harvested, so part of the benefit comes back as tax later. We priced that trade in what “tax alpha” actually means and what $121,281 of harvested losses was worth.

Harvesting and the write-off rationalization both involve a realized loss. What separates them is the order of the decisions. The harvester took a market decline they could not avoid and asked what it was worth at tax time. The rationalizer took on, or held onto, a loss and pointed to the tax treatment as the reason it was acceptable.

Decide without the tax, then add it back

Make the decision as if taxes did not exist, then bring taxes in and see whether they change it.

• For a position that is down: if this loss produced no deduction, would I still want to own it? If not, sell, and the deduction is a partial offset. If yes, and you hold it in a taxable account, harvesting into a similar fund may capture the loss without changing what you own. The question that matters is what you expect from here, and the price you paid belongs only in the tax calculation.
• For a new investment or expense: would I do this if it were not deductible? If the answer is no, the investment is being justified by t × L, and t × L is always less than L. The case for going ahead has to rest on something else.
• For a loss you are sitting on: work out how fast you can use it. With no gains coming, a large carryforward at $3,000 a year is worth well under its face value, and gains you were going to realize anyway are the fastest way to use it.

When the tax treatment should change the decision

Taxes can legitimately tip a decision, and the tax ledger and the economic one can disagree in both directions.

• Depreciation on an asset that is not losing value. A rental can show a tax loss while producing cash and appreciating. That is a gap between the tax ledger and the economic one, in your favor, subject to the passive loss rules and recapture at sale. The rental write-off guide works through it.
• Credits and subsidies. A refundable credit is worth its face value, and some incentives are large enough to turn a marginal project into a good one. That is an after-tax return calculation on the whole investment, done before you commit.
• A year with large gains. If you are realizing large short-term gains, a harvested loss is worth close to your full marginal rate immediately, and the $3,000 limit does not bind.
• Small-business stock that qualifies under section 1244. Ordinary-loss treatment up to $50,000 or $100,000 a year changes the after-tax downside of a direct startup investment. It does not change the odds that the startup fails.
• Inflation. The tax code measures gains and losses in nominal dollars. A position that is flat in nominal terms after a decade has lost purchasing power and produces no tax loss at all. Tax accounting and economic results can disagree in either direction.

What we recommend

• Keep score in after-tax net worth. Taxes paid is a cost to minimize only when the pre-tax outcome is held fixed.
• Value a capital loss at what you can use. Offset against which gains, at which rate, and how many years at $3,000 before it runs out or you do.
• Harvest losses the market gives you. In a taxable account, selling a declined fund and buying a similar one is cheap and often worth doing. Avoid your IRA for the replacement purchase.
• Put no weight on the deduction when deciding to buy. An investment has to justify itself before tax. The deduction then makes a sound decision a little cheaper.

How Summitward helps

What a loss is worth depends on your marginal rate and whether you have gains to absorb it. Tax Projection computes projected AGI and your marginal federal rate from your own income, and Tax-Loss Harvesting splits a harvest from your own tax lots into the part that offsets this year’s gains, the $3,000 against ordinary income, and the carryforward.

Frequently asked questions

Is it ever worth losing money for the tax deduction?

Not when the tax loss equals the economic loss, which is the normal case for an individual investor. The deduction is worth at most your marginal rate times the loss, so you keep the rest of the loss. Buying losses paid only when the tax loss was larger than the economic one, as in the tax shelters of the early 1980s.

How much of a capital loss can I deduct each year?

Capital losses offset capital gains in full. Beyond your gains, you can deduct up to $3,000 a year against ordinary income ($1,500 if married filing separately), and carry the rest forward for as long as you live. The carryforward cannot be used by your estate or heirs.

Does a big loss mean the government pays for part of it?

Only in proportion to how quickly you can use it. With gains to offset, the government absorbs your capital gains rate on the loss. With no gains, it absorbs your ordinary rate on $3,000 a year. For one modeled investor in the 37% bracket with no gains, that came to $18,725 in present value on a $100,000 loss.

Can I deduct a loss on my house or car?

No. Losses on property held for personal use are not deductible, apart from casualty and theft losses. A rental property or a business asset is treated differently.

Should I keep a losing stock so I can deduct the loss later?

Holding does not make the deduction larger; it only moves the loss up or down with the price. If you no longer want the stock, selling now realizes the loss now. If you do want the exposure, you can harvest by selling and buying something similar that is not substantially identical. Either way, what you expect from the position from here should decide it, and the tax loss should not.

Is tax-loss harvesting just rationalizing losses too?

No. Harvesting takes a decline that already happened, in something you still want to own, and turns it into a tax asset. It does not create losses or change what you own. It does lower your basis, so part of the benefit is a deferral that comes back when you eventually sell.

Key takeaways

• A deduction returns your marginal rate, at most. Losing $10,000 to save $2,400 of tax leaves you $7,600 poorer.
• The comparison is loss against no loss. Tax treatment changes how large a loss is, and for an ordinary investment loss it cannot turn the loss into a gain.
• The $3,000 limit shrinks large capital losses with no gains to offset. For one modeled investor in the 37% bracket, a $100,000 loss was worth $18,725 in present value at a 5% discount rate, and unused carryforwards expire at death.
• Some losses produce nothing. Personal-use property, hobbies, and wash sales into an IRA get no deduction.
• Buying losses worked only when tax losses exceeded economic ones. Pre-1986 shelters combined a 50% top rate, a 60% capital gains exclusion, and partnership losses that offset wages.
• Harvesting is fine because the loss came first. Decide what you want to own without the tax, then use the tax rules on the result.

Related guides

• The Beach Rental That Cut Taxes and Lost Money follows one real vacation rental from purchase to the sale-year release of fifteen years of suspended losses.
• What a $16,364 Rental Write-Off Is Actually Worth separates the tax, cash, and economic ledgers for a rental property.
• Yes, an Unrealized Loss Is Still a Loss covers the disposition effect and a decision framework for a position that is down.
• What $121,281 of Harvested Losses Was Worth prices one investor’s harvest after the $3,000 limit and the lower basis.
• What “Tax Alpha” Actually Means explains why the popular 1% figure depends on how you eventually exit.
• Tax-Loss Harvesting walks through the mechanics, the wash-sale rule, and replacement funds.

Sources

1. Internal Revenue Service, “Credits and Deductions for Individuals.” irs.gov
2. Internal Revenue Service, Publication 550, Investment Income and Expenses (2025), sections on capital losses, the decedent’s capital loss, section 1244 stock, and worthless securities. irs.gov
3. 26 U.S.C. § 1211(b), limitation on capital losses for individuals. law.cornell.edu
4. Internal Revenue Service, Revenue Procedure 2025-32, inflation adjustments for tax year 2026. It adjusts brackets and many other amounts and does not adjust the section 1211(b) limit. irs.gov
5. Internal Revenue Service, Topic No. 409, “Capital Gains and Losses.” irs.gov
6. 26 U.S.C. § 165(c), limitation on losses of individuals. law.cornell.edu
7. 26 U.S.C. § 183, activities not engaged in for profit. law.cornell.edu
8. 26 U.S.C. § 67(h), suspension of miscellaneous itemized deductions for taxable years beginning after 2017, as amended by Public Law 119-21 (2025), which removed the 2025 sunset. law.cornell.edu
9. Internal Revenue Service, Revenue Ruling 2008-5. irs.gov
10. Andrew A. Samwick, “Tax Shelters and Passive Losses after the Tax Reform Act of 1986,” in Martin Feldstein and James Poterba, eds., Empirical Foundations of Household Taxation (University of Chicago Press, 1996), 193–233. The 40% figure is Samwick citing Petska (1992); partnership sales are Robert A. Stanger & Co. tabulations of SEC filings. nber.org
11. Susan Nelson and Tom Petska, “Partnerships, Passive Losses, and Tax Reform,” IRS Statistics of Income. Loss figures are for limited partnerships with losses, in ordinary income, 1981–1987. irs.gov
12. Zoe Callaway, William McBride, Yihan Chen, and Nicolo Pastrone, “US Tax Literacy Poll,” Tax Foundation, December 18, 2024. Survey conducted by Public Policy Polling; question Q18. taxfoundation.org
13. Richard H. Thaler, “Mental Accounting Matters,” Journal of Behavioral Decision Making 12, no. 3 (1999): 183–206. doi.org
14. Hal R. Arkes and Catherine Blumer, “The Psychology of Sunk Cost,” Organizational Behavior and Human Decision Processes 35, no. 1 (1985): 124–140. doi.org
15. Terrance Odean, “Are Investors Reluctant to Realize Their Losses?” Journal of Finance 53, no. 5 (1998): 1775–1798. Trades from 10,000 discount-brokerage accounts, 1987–1993. The 3.4-point gap is the average excess return over the next 252 trading days of winners sold minus losers held. doi.org
16. George M. Constantinides, “Optimal Stock Trading with Personal Taxes: Implications for Prices and the Abnormal January Returns,” Journal of Financial Economics 13, no. 1 (1984): 65–89. nber.org
17. Zoran Ivković, James Poterba, and Scott Weisbenner, “Tax-Motivated Trading by Individual Investors,” American Economic Review 95, no. 5 (2005): 1605–1630. nber.org
18. Shomesh E. Chaudhuri, Terence C. Burnham, and Andrew W. Lo, “An Empirical Evaluation of Tax-Loss-Harvesting Alpha,” Financial Analysts Journal 76, no. 3 (2020): 99–108. CRSP monthly data, 500 largest stocks, 1926–2018. dspace.mit.edu
19. Kevin Khang, Thomas Paradise, and Joel Dickson, “Tax-Loss Harvesting: An Individual Investor’s Perspective,” Financial Analysts Journal 77, no. 4 (2021): 128–150. doi.org

Editor’s note

Educational content, not tax advice. Federal rules only; state treatment of capital losses differs. Rules were checked against the IRS publications and statutes listed above in September 2026. The $100,000 example is one modeled investor under the stated assumptions; your rate, gains, and time horizon will produce different figures.