Tax-Loss Harvesting: Turn Bond Losses Into Tax Savings

I opened a brokerage statement last month and did the thing most of us do: I scrolled past the bond sleeve because equities were the loud part of the page. Then I stopped. A fund I had bought when yields felt boring was sitting well below what I paid. Not a disaster. Not a reason to abandon fixed income. Just a quiet red number that, in a taxable account, can sometimes be more useful than a green one. That is the odd little gift of a rising-rate stretch. Prices fall. Losses appear on paper. And if you know how to handle them, some of those losses can shrink a tax bill you were going to pay anyway.

Higher Treasury yields have pushed a lot of bond prices lower. Investors who bought individual bonds, bond mutual funds, or exchange-traded funds when yields were softer may now be holding unrealized losses. Selling in a taxable brokerage account can turn those paper losses into real capital losses, which can offset taxable gains elsewhere. The strategy has a plain name: tax-loss harvesting. It is not a trick, and it is not free money. Done carelessly, it can even backfire. Done with a clear head, it is one of the few times a disappointing price can work in your favor.

Perhaps the most interesting aspect is timing. Plenty of people wait until the last two weeks of December, then scramble. By then spreads are wider, replacements are harder to think through, and the wash-sale clock gets messy if you are also rebalancing a year-end bonus into the same sleeve. Looking now, while yields are still elevated and prices are still soft, gives you room to be picky.

Why Bond Losses Can Quietly Cut a Tax Bill

Bonds and stocks do not fall for the same reasons, which is why this moment feels different from a plain equity dip. When yields rise, existing bonds with lower coupons become less attractive. Buyers demand a discount so the yield to maturity matches what the market is offering on new paper. That discount is the loss you see. Nothing about the issuer has to be broken. The math just moved.

Longer maturities feel it more. A small move in the 30-year yield can knock a surprising amount off price. Intermediate funds move less, short funds less still, and cash-like holdings barely twitch. If your statement shows a deep red line on a long Treasury fund and a shrug on a short-term fund, that is not a bug. Duration did what duration does.

This week, longer Treasury yields reached levels not seen in roughly a generation, territory last visited in the early 2000s, before easing after a weaker-than-expected jobs report. I am not going to pretend anyone can call the next print. What matters for tax planning is simpler. A lot of purchases made in the low-yield years are now underwater. Those purchases live in taxable accounts for plenty of households, because the bond sleeve was the “safe” bucket and nobody thought to ask whether the account type matched the strategy.

A paper loss is a mood. A realized loss is a line on a tax form. Only the second one can offset a gain.

Wealth advisors who also hold tax credentials tend to put it bluntly. Disposing of bonds that are underwater can convert unrealized losses into capital losses you can use against other capital gains. That is the whole engine. Everything else is guardrails.

What Harvesting Actually Does, Without the Jargon Fog

You sell an investment for less than its tax cost basis. The difference is a capital loss. That loss nets against capital gains in the same year. If losses win the netting fight, a limited slice can reduce ordinary income. Whatever is left generally carries forward. That is the skeleton. The flesh is account type, holding period, what you buy next, and whether the trade still belongs in the portfolio you claim to want.

Short-term and long-term do not mix in a free-for-all. The code nets short-term gains and losses together, long-term gains and losses together, then nets the two results. A long-term loss can still offset a short-term gain, but the order matters for the rate you save. Offsetting a short-term gain, which is taxed like wages, is usually the richer prize. Offsetting a long-term gain taxed at preferential rates still helps, just less per dollar. I have found that people fixate on the size of the loss and ignore which gain it is eating. The size is not the whole story.

Harvesting also does not create a deduction out of thin air if you immediately rebuild the exact same position in a way the rules dislike. More on that in a minute. The point here is mechanical. Sell below basis. Book the loss. Use the loss. Stay invested if staying invested was the plan.

Where the Strategy Works, and Where It Is a Waste of Clicks

Tax-loss harvesting generally does nothing useful inside a 401(k), traditional IRA, or Roth IRA. Those accounts do not recognize capital gains or losses the way a brokerage account does. Selling at a loss in a retirement plan does not hand you a deduction on this year’s return. It just resets the holding inside a wrapper that was already tax-deferred or tax-free.

The opportunity lives in taxable brokerage accounts. Joint accounts, individual accounts, trust accounts that file their own returns. If the bond lot is in one of those, and the market value is below what you paid, you have a candidate. If it is inside a workplace plan, leave it alone for tax reasons and judge it only as an investment.

• Taxable brokerage account: harvesting can offset gains and, within limits, ordinary income.
• Traditional IRA or 401(k): a sale at a loss does not create a deductible capital loss.
• Roth IRA: same idea, and a wash-sale purchase here can still poison a loss taken elsewhere.
• Health savings account used as an investment sleeve: treat it like a retirement wrapper, not a harvesting pool.

One wrinkle people miss: the wash-sale rule can reach into an IRA even when the loss was taken in a taxable account. Buying a substantially identical bond fund inside a Roth within the window can disallow the loss you thought you booked. The accounts are separate for contribution rules. They are not separate for this particular trap.

A Four-Step Pass Before You Touch a Trade Ticket

Experts who do this for a living tend to walk clients through the same short list. I like it because it stops the “I saw a red number, so I sold” reflex.

1. Find the losses that actually sit in taxable accounts. Ignore the red ink in retirement plans for this exercise.
2. Sell only what is trading below its tax cost basis. A sad chart is not a loss. A trade confirmation is.
3. Match the loss to real gains, then to the ordinary-income allowance if losses still remain, and track anything that carries forward.
4. Reinvest with intention. The goal is to keep the risk you wanted, not to sit in cash hoping for a better entry you cannot time.

That fourth step is where good plans go soft. Someone harvests a broad bond fund, parks the cash “for a few weeks,” and then rates drop, prices jump, and the replacement costs more than the thing they sold. The tax benefit was real. The opportunity cost was larger. Staying invested, in something that is not substantially identical, is usually the adult move.

Cost Basis Is the Number That Decides Everything

Your loss is not “what the fund was worth at the peak.” It is market value minus tax cost basis. Basis is what you paid, adjusted for reinvested dividends, return of capital, and certain corporate actions. Bond funds reinvest often. Those reinvested distributions buy new shares at new prices. Some lots may be underwater. Some lots bought last month may not be.

Specific-lot identification beats average cost when you are harvesting. If your custodian defaults to average cost for a mutual fund, you can end up selling a blend that dilutes the loss or, worse, realizes a gain on newer shares. Check the lot view before you click. ETFs are usually easier here because basis is tracked share by share, but defaults still vary. I have watched people “harvest” a position and discover the 1099 showed a small gain because the method ate the old lots and spat out the new ones.

Individual bonds have their own wrinkles. You might have bought at a premium or a discount. Premium amortization and market discount rules can change the character of what you think is a capital loss. A bond bought above par and held while rates rose may show a price drop that is partly the premium you were already writing down. That is a conversation for a tax preparer, not a guess on a Sunday night. The headline idea still holds. If the amount realized is below adjusted basis, you may have a capital loss. Confirm the adjusted part.

How the Offsets Actually Stack

Imagine you sold stock earlier this year for a long-term gain, and a trading position for a short-term gain. Later you sell a bond fund at a loss. The loss does not wander around the return looking for the nicest place to sit. It follows netting rules. Short with short. Long with long. Then the leftovers meet.

If your bond loss is long-term and your biggest gain is short-term, the long-term loss can still reduce that short-term gain after the same-character netting is done. That is often excellent, because short-term gains are taxed at ordinary rates. A dollar of loss that knocks out a dollar of short-term gain can be worth more than a dollar of loss that knocks out a long-term gain taxed at 15 percent. State taxes pile on top in many places, which makes the gap wider.

There is also the net investment income tax for higher earners, that extra 3.8 percent on certain investment income above a threshold. Capital gains can feed that calculation. Reducing gains can reduce that extra layer too. It is not the reason to harvest, but it is a reason not to shrug at a modest loss if your income is already near the line.

Married taxpayers filing separately generally get half that ordinary-income allowance, $1,500 rather than $3,000. The carryforward does not vanish at year-end. It waits. People forget they have one. Pull last year’s Schedule D before you assume this year’s harvest is starting from zero.

The $3,000 Rule, and Why It Is Not the Main Event

If capital losses exceed capital gains, you may generally use up to $3,000 of the net capital loss to reduce ordinary income for the year. That number has been stuck for a long time. Inflation did not get a vote. For a household with a large equity gain, the $3,000 is a footnote. The real money is the gain you no longer recognize.

For a household with no gains at all, the $3,000 is the whole prize this year, and the rest rides forward. That can still be worth doing if the position no longer fits, or if you want the loss banked for a future sale of a business, a rental, or a concentrated stock lot. It is a weaker reason to churn a bond fund you were happy to hold. Transaction costs, bid-ask spreads on less liquid funds, and the risk of a bad replacement can eat a deduction that only saves you your marginal rate on three thousand dollars.

Run the arithmetic before you fall in love with the idea. A $3,000 deduction at a 24 percent federal rate is $720, before state. If the round trip costs you more than that in spread and a missed coupon, you harvested a feeling.

Carryforwards Are a Quiet Asset

Unused capital losses generally carry forward indefinitely, until they are used up. They keep their character in a way that still follows the netting rules each year. A large harvest in a dull year can shelter a liquidity event two or three years out. I have seen families treat a loss carryforward like an emergency fund for the tax return. It is not cash. It is a coupon against future gains. Still, knowing it is there changes how you stage a sale.

Death does not pass a capital-loss carryforward to heirs in the simple way people hope. That is a grim footnote, and it matters for older investors sitting on a huge unused loss they keep “saving.” Using some of it while you can is sometimes kinder than hoarding it. This is not advice for a specific estate. It is a nudge to ask the question.

The Wash-Sale Rule Is the Part That Bites

Sell a bond, a bond fund, or an ETF at a loss, and the tax loss can be disallowed if you buy the same or a substantially identical security within 30 days before or after the sale. The window is 61 days wide if you count both sides. A purchase the week before the sale counts. A purchase the week after counts. An automatic dividend reinvestment can count. A purchase in an IRA can count.

Substantially identical is clearer for a single stock than for a bond fund. Selling one broad aggregate bond ETF and buying another broad aggregate bond ETF from a different sponsor can be uncomfortably close, especially if the holdings and duration line up. Selling a long Treasury fund and buying a short corporate fund is much easier to defend. The line in the middle is where people get sloppy. Portfolio managers at large asset managers will tell you that leaving fixed income entirely is not required. Moving to a different fund that still fits the goal, without being substantially identical, is often enough. Making that call with confidence is harder than the brochure suggests. A tax professional who will actually look at the two prospectuses is worth the hour.

The wash-sale rule does not ban you from owning bonds. It bans you from pretending you left a position you never really left.

– A framing I keep taped above the trade screen

If the loss is disallowed, it is not deleted from the universe. It is generally added to the basis of the replacement shares. You may get it back later, when you sell the new lot. That is cold comfort if you needed the loss this year to offset a gain you already booked. And if the replacement was bought inside an IRA, the basis adjustment can disappear into an account that does not use basis the same way. That version of the mistake is the expensive one. Review every account, including a spouse’s, before you sell.

What “Different Enough” Can Look Like

I am not going to hand you a pairing chart and call it safe. The standard is facts and circumstances, and I am not your examiner. What I will say is how thoughtful investors tend to create distance.

• Change duration in a real way. Long government to intermediate credit is a different bet, not a costume change.
• Change issuer mix. Treasuries to municipals is a different risk and a different tax character. Be sure you want that character.
• Change structure. An individual bond ladder is not a total-market bond ETF, even if both are “fixed income.”
• Wait out the window in a genuinely different holding, then return, if returning was the plan. Thirty days is not a market-timing strategy. It is a rule.

Municipal bonds deserve a separate caution. Losses on munis can still be capital losses, which is useful. The income was tax-favored, and selling can change what your after-tax yield looks like going forward. Harvesting a muni fund into a taxable Treasury fund because the Treasury fund “feels safer” can raise your annual tax bill even as it banks a loss. Run both sides.

Individual Bonds, Funds, and the Ladder That Should Often Stay Put

Not every red bond should be sold. If you built a ladder of individual bonds, you like the income, and you plan to hold to maturity, the price on the statement is mostly a mark. Barring default, you expect par back at maturity. Selling locks in the discount and forces you to reinvest at whatever the market is offering, which may be fine, and may also be a fuss you do not need. Advisors who work with ladder clients often say exactly that. As bonds mature, roll the proceeds into newer, higher-yielding issues. The higher yield arrives without a tax-loss project.

Funds are different. A bond fund does not mature. It constantly replaces holdings. The loss is not going to “pull to par” in the clean way a single bond does, because the portfolio keeps extending. Waiting for a fund to recover is a bet on yields, not a contractual path back to your cost. That makes funds cleaner harvesting candidates, and also easier to mishandle if you swap into a twin.

There is a middle case. You own individual bonds you no longer want, credit you have soured on, or a maturity that no longer matches a goal. The tax loss is a sweetener, not the thesis. Sell because the bond is wrong. Enjoy the loss because it showed up.

A simple hold-or-harvest filter:
  Need the cash flow and will hold to maturity? Often hold.
  Fund loss, taxable account, gains elsewhere? Often harvest.
  Selling only to feel clever? Stop.

Rebalancing Is the Grown-Up Reason to Be Here

Yields at today’s levels are not just a tax story. They are a portfolio story. For years, the bond sleeve paid so little that people drifted into longer duration, narrower credit, or simply a smaller allocation because stocks were doing the emotional work. Higher yields let you rebuild income without reaching as far. A portfolio manager at a global asset manager put the combination cleanly: at these yield levels, investors can look at tax-loss harvesting and at rebalancing in the same sitting.

That review should include time horizon, income needs, interest-rate risk, and diversification. A drop in one sector is a decent excuse to revisit performance, quality, and yield across the board. I agree with that, with one caveat. A drop is also a decent excuse to overtrade. Write down the allocation you wanted before you open the loss report. If the harvest pulls you toward that allocation, good. If it pulls you toward whatever is easiest to justify as “not substantially identical,” you are letting the tax tail wag the portfolio.

Transaction costs, tax consequences, and any change in risk belong in the same note. A tax loss can be valuable. It should support the strategy, not dictate it. That sentence is worth keeping. I have broken it before, usually in December, usually while telling myself I would fix the allocation in January.

Age, Horizon, and Who Can Afford the Volatility

Someone in their thirties can ride a rough patch in bonds more easily than someone drawing income next year. Retirement specialists make this point often, and it lands. A younger investor harvesting a long bond fund and replacing it with something similar in risk is mostly doing tax housekeeping. A retiree harvesting the same fund and replacing it with a longer fund, because the longer fund was the available “different” option, may have just increased the chance of another drawdown right when withdrawals start.

Sequence matters more near retirement. A loss you harvest is fine. A risk you accidentally upgrade is not. If the bond sleeve is there to fund five years of spending, do not let a wash-sale workaround stretch the maturity. Park the economic exposure where the spending plan already said it should live. The tax form is secondary to the cash-flow calendar.

There is a kinder version for retirees too. Higher yields mean new bonds and new fund purchases can throw off more income than the old book. Harvesting a tired lot and rebuilding at today’s yield can raise the paycheck, as long as credit quality and duration stay inside the lines you set. That is rebalancing with a tax assist. It is not a dare.

A Worked Sketch, So the Dollars Feel Real

Say you bought $80,000 of an intermediate bond fund a few years ago. It is worth $68,000. You also sold stock this spring for a $20,000 long-term gain, and you have no other gains or losses. You sell the fund. You realize a $12,000 long-term loss. That loss nets against the $20,000 gain. You recognize $8,000 of long-term gain instead of $20,000. At a 15 percent long-term rate, the federal tax on that gain drops by $1,800, before state and before any net investment income tax. You then buy a different fixed-income fund with a clearly different duration and sector mix, on the same day, because you still want bonds.

Change the facts. No stock sale this year. The $12,000 loss has nothing to offset except the $3,000 ordinary-income allowance. You deduct $3,000. You carry forward $9,000. Still useful if a gain is coming. Thin if you paid a wide spread to do it and the replacement drifts.

Change them again. You sold the fund, then your dividend reinvestment on a nearly identical fund in another account bought $400 of shares nine days later. A slice of the loss may be disallowed. The rest stands. This is why the boring account audit beats the clever trade.

Rough federal sketch: loss used against long-term gain × 15% = tax deferred or saved this year. State tax and the 3.8% surtax can raise the value. Spreads and a bad swap can erase it.

Mistakes That Keep Showing Up

The patterns are boring, which is why they persist. People harvest inside an IRA and expect a deduction. People sell and buy the same ticker in a spouse’s account. People forget that a purchase 20 days before the sale is inside the window. People use average cost and wonder where the loss went. People harvest a muni loss and replace it with a taxable bond, then act surprised in April. People sit in cash for the wash-sale window and miss a rally that was worth more than the deduction.

Another one, quieter: harvesting losses while also donating appreciated stock. Those two moves can complement each other. You donate the low-basis shares, avoid the gain, and harvest losses elsewhere to mop up gains you could not avoid. They can also be coordinated badly if you donate the lot you meant to harvest. Look at the lots. Label them. Future you, doing taxes, is not as sharp as tonight’s you.

And then there is the performance chase dressed up as tax work. A sector fund is down, so it gets sold “for the loss,” and the replacement is whatever did well this quarter. That is not harvesting. That is regret, with a receipt.

Questions Worth Asking Before the Order Goes In

Sit with these. If you cannot answer them, you are not ready to click.

1. Is this lot in a taxable account, and is market value below adjusted basis on the specific shares I will sell?
2. What gains will this loss actually offset, and at what rate?
3. Do I already have a carryforward that changes the math?
4. What will I buy, and can I explain why it is not substantially identical?
5. Have I checked every account, including IRAs and a spouse’s accounts, for purchases inside the 30-day window on either side?
6. Does the replacement keep my duration, credit, and income plan intact?
7. After costs, is the tax benefit still larger than the hassle?

If you hold bonds to maturity and the income is the point, question five and six may end the exercise. Good. Not harvesting is a decision. It is often the right one.

Yield, Quality, and the Part That Is Not About April

Anytime a corner of the market sags, it is a fair moment to compare yield and quality instead of staring at the loss column. A Treasury fund and a high-yield fund can both be red. They are not the same repair job. One is rate risk. The other may be credit risk wearing a bond costume. Harvesting the second into more of the second, because you want the loss and you like the yield, can concentrate the exact risk that produced the loss.

I prefer to separate the questions. First, do I still want this risk. Second, is there a tax benefit to changing the lots. If the answer to the first is no, sell. The loss is a passenger. If the answer to the first is yes, only swap if you can keep the risk and still clear the wash-sale smell. If you cannot, hold, and revisit when a gain elsewhere makes the friction worth it.

Higher starting yields also change the recovery math. A bond fund yielding much more than it did when you bought it can claw back a price decline through income, slowly, if yields stop rising. That does not erase the case for harvesting. It does mean the “it will never come back” story is lazier than it sounds. Income is part of the total return you were supposed to be measuring all along.

Couples, Joint Accounts, and the Unsexy Coordination Problem

Joint taxable accounts make harvesting look simple until two people are buying. One spouse sells a total bond fund on Tuesday. The other, on autopilot, buys the same fund on Friday inside a Roth because the monthly contribution landed. The loss is wounded. Nobody meant to do it. The fix is a shared watchlist and a 61-day pause on that exposure across the household.

Filing status changes the ordinary-income cap, as noted. It does not change the wisdom of the trade. What it changes is the conversation. If you file separately, do not assume the household gets two full $3,000 allowances stacked in the way a casual article implies. Read the instruction for your status, or ask the person who signs the return.

Community-property states add another layer to basis and to who is treated as buying. I am not going to fake precision there. If you live in one, the household audit matters even more.

A Calendar That Beats the December Scramble

Review taxable fixed income when yields have already done the damage, not when the holiday inbox is full. A mid-year or early-autumn pass leaves time to pick a replacement, turn off dividend reinvestment on the old fund, and let the wash-sale window close before you even think about year-end gain taking.

Some investors harvest twice a year and call it done. Others check after any sharp yield move. Either habit beats a single panicked afternoon. The market does not owe you a December dip. This year’s bond weakness showed up earlier. Using it is allowed.

Keep a one-page log. Date sold, ticker, lot basis, proceeds, replacement ticker, why it is different, accounts checked. Future you will not remember. The log will. It also makes the preparer’s life shorter, which sometimes shows up in the bill.

What I Would Not Do With This Window

I would not sell a ladder I built on purpose just to manufacture a loss. I would not swap into a fund I cannot explain, solely because a blog said the names were different enough. I would not harvest in a retirement account and expect a party on the return. I would not ignore state tax, because in a high-tax state the deduction is meatier and the muni decision is sharper. I would not let a tax idea raise the risk of money I need inside five years.

I would sell a taxable bond fund that is below basis, if I have gains to offset, if the replacement keeps my role for bonds intact, and if I have swept the household for look-alike buys. That is a narrow door. Narrow doors are how you avoid clever mistakes.

The broader market can still shove yields around after a soft jobs print or a hot one. Friday’s retreat does not cancel the losses created on the way up. It does remind you that waiting for the perfect price is a different hobby from tax planning. If the position is wrong, or the loss is useful and the swap is clean, the perfect week is the week you understand the trade.

Putting the Pieces Back on One Page

Bond prices fell because yields rose. That left taxable investors with losses they did not have to realize, and can realize if they want. Realizing them can offset capital gains, chip away at ordinary income up to a small annual cap, and stock a carryforward. The wash-sale rule polices the next purchase, including purchases in IRAs, for 30 days on either side. Ladders held to maturity are often better left alone. Funds in taxable accounts are the usual candidates. Rebalancing at higher yields can be the real prize, with the tax loss riding along.

None of this replaces a look at your own basis, your own gains, and your own replacement. Rules in this area are specific, and a generic sketch is not a ruling. What the sketch can do is stop the two failure modes I see most: ignoring a useful loss until the year is gone, and grabbing a loss in a way that the rules simply hand back.

If the red number on the bond line has been annoying you, give it a job. Or leave it alone, on purpose, because the income and the maturity still match the life you are funding. Both are grown-up answers. Only one of them should be an accident.