Capital Loss Carryover: The $3,000 Rule and Beyond

A realized capital loss is not just a bad year; it is a carryable tax attribute that can shelter future gains for as long as you live. The mechanics are simple but rigid: losses net against gains first, then trickle out against ordinary income at $3,000 per year, and the rest rolls forward on Schedule D.

The $3,000 figure is the piece everyone knows and the piece that matters least. It has been fixed since 1978 and is not indexed. For an investor with a six-figure carryover, the real value is not the $3,000 drip; it is the unlimited offset against future gains, which turns the carryover into a free pass for a future rebalancing, property sale, or concentrated stock unwind.

This page covers the netting order, the carryover math, and the coordination points with tax-loss harvesting, which has its own dedicated guide on this site.

Step one happens inside each bucket: short-term losses net against short-term gains, and long-term losses net against long-term gains. Step two nets the buckets against each other if one is negative and the other positive.

If the combined result is a net loss, up to $3,000 deducts against ordinary income this year ($1,500 for married filing separately), and the remainder carries forward. The carryover keeps its character: a short-term carryover arrives next year as a short-term loss, first netting against short-term gains, which are the expensive kind (taxed at ordinary rates up to 37%).

That character rule creates a quiet preference order: short-term losses are more valuable because they shelter ordinary-rate gains first. Long-term losses that end up offsetting 0% or 15% long-term gains do less work per dollar.

One boundary worth marking: these rules govern capital assets. Losses on personal-use property (your home, a car) are never deductible, and ordinary losses (business bad debts, certain small-business stock) follow different rules entirely. The carryover regime is for investment assets: securities, crypto, investment real estate, and the like, where character and holding period drive everything.

Two mechanics inside this math are worth pausing on. The $3,000 comes off ordinary income at your top marginal rate, so its cash value scales with your bracket: $960 at 32%, $1,110 at 37%. And the netting happened before the limit applied: the $25,000 of gains absorbed losses first, which is why only $15,000 was left for the ordinary-income offset and the carryover. Reversing that order in your head is the most common way people misestimate what a loss year was worth.

Tax-loss harvesting (covered in depth in our dedicated guide) generates the losses; the carryover rules decide what they are worth. If you already carry a large loss balance and expect no significant gains, harvesting more losses adds value at only $3,000 per year, which at a 35% rate is about $1,050 of annual tax savings. The marginal harvest may not justify transaction costs and wash-sale management.

The calculus flips when gains are coming: an upcoming property sale, a planned exit from concentrated stock, or annual mutual fund distributions. Losses banked ahead of known gains offset at the gain’s full rate, including the 3.8% NIIT for high earners, making pre-gain harvesting worth up to 23.8 cents or more per dollar.

Mind the wash-sale rule when harvesting securities: repurchasing a substantially identical security within 30 days before or after the sale defers the loss. Directly held crypto is currently outside the wash-sale statute, a difference our crypto wash-sale guide covers.

Mutual funds in taxable accounts add a passive angle: funds distribute their internal realized gains to shareholders late each year, and those distributions are offsettable gains like any other. An investor with a standing carryover can hold gain-heavy funds with less tax drag, while an investor without one may prefer ETFs precisely to avoid distributions the carryover would otherwise have absorbed.

Death extinguishes carryovers. A loss that is not used on the final return (or the survivor’s share on a joint return) is gone; heirs cannot inherit it, and appreciated assets they receive get a stepped-up basis anyway. Large carryovers late in life argue for realizing gains, not deferring them.

Divorce splits carryovers according to whose losses they were. Software imports lose them: a new preparer or a switch of tax software that drops the carryover worksheet forfeits real money silently, and amended or corrected 1099s can change the number after the fact.

And the $3,000 ordinary offset is not elective. The deduction applies whether or not it saves tax, so in a zero-income year part of the carryover can be absorbed with no benefit; the worksheet handles this, which is one more reason to run it every year rather than assuming the balance rolled intact.

Married couples should also know the filing-status wrinkle: on a joint return the losses of either spouse offset the gains of either spouse, but if you later file separately, each spouse keeps only their own losses, and the annual ordinary-income allowance drops to $1,500 each. Carryovers built jointly need tracing back to whose assets generated them.

The payoff scenario deserves its own walkthrough, because it is where a carryover stops being a $3,000 trickle and becomes real money. Suppose the investor from the example above carries the $12,000 loss into a year with a planned $100,000 long-term gain from a property or concentrated stock sale, with income high enough to face the 15% capital gains rate plus the 3.8% NIIT.

The carryover offsets the first $12,000 of the gain before any tax is computed. At an 18.8% combined rate, that is roughly $2,256 of federal tax avoided in one year, against the $960 the same loss would have produced drip-fed against ordinary income at $3,000 per year over four years. Timing the gain into the carryover, or the carryover into the gain, is the whole game.

This is also the argument for checking the carryover before agreeing to an installment sale, an opportunity-zone deferral, or any structure that pushes gain into future years. Deferral strategies and loss carryovers compete for the same gains; sequencing them badly can strand the loss in years with nothing to offset.