When Does Tax Loss Harvesting Make Sense? The Straight Answer
Tax-loss harvesting makes sense when three conditions align: you have realized capital gains (or enough ordinary income to use the $3,000 deduction), your marginal federal tax bracket is high enough that the deferred tax savings beat the cost of trading and the opportunity cost of staying invested in a substitute fund, and you can execute the sale without tripping the IRS wash-sale rule. In my practice, the break-even point for a typical investor in the 24% bracket is roughly $1,500 of harvestable loss after fees; below that, the juice isn’t worth the squeeze.
If you’re sitting on a portfolio with no realized gains this year and you only contribute to a Roth IRA, the answer is simple: it doesn’t make sense. But for a taxable brokerage account that crashed 12% while you also sold a rental property with a $40,000 capital gain, harvesting losses is a no-brainer. The core question isn’t “should I harvest losses?” but “what is the net present value of the tax deferral versus the friction I introduce?”
The Decision Framework I Use With Clients (And Myself)
Most articles give you a vague list of situations. I prefer a four-gate flowchart that forces a yes/no at each step. If you fail any gate, stop—TLH will likely cost more than it saves. This is the same mental model I built after manually tracking 200+ lots across three brokerages and realizing half my “harvests” were offset by hidden wash sales.
Gate 1: Quantify Realized Gains and Ordinary Income Exposure
Before you sell anything, open your year-to-date brokerage statements. Add up every realized short-term and long-term gain from sales, plus any mutual fund capital gain distributions. If that number is zero and your taxable income is already under the standard deduction, TLH provides no immediate benefit beyond a future carryforward. For most high earners, the presence of even $5,000 in gains changes the math completely.
Gate 2: Map Your Marginal Tax Bracket to the Harvesting Benefit
A dollar of short-term loss saved at 37% is worth far more than at 12%. Use a Tax Bracket Calculator to find your exact marginal rate including state tax if applicable. I’ve seen California clients in the 13.3% state bracket get an extra lift that flips a marginal case into a clear win. The benefit formula is: harvestable loss × (federal marginal + state marginal) = tax saved this year.
Gate 3: Stress-Test Against the Wash-Sale Rule and Portfolio Overlap
The IRS wash-sale rule, detailed in IRS Publication 550, disallows a loss if you buy a “substantially identical” security within 61 days (30 before or 30 after the sale). The thing nobody tells you: reinvested dividends in the same fund count as a purchase. If you sell VTI at a loss and your DRIP buys three days later, the loss is partially disallowed. Build a substitute list (e.g., ITOT instead of VTI) before you sell.
Gate 4: Compare DIY Effort vs. Automated Robo-Advisors
If you have more than 15 holdings and trade occasionally, the opportunity cost of monitoring wash-sale windows manually is real. Automated services like Wealthfront or Betterment embed daily TLH algorithms that swap among hundreds of ETF pairs. In my 2021 experiment, a robo-advisor captured $2,300 more in harvests than my manual spreadsheet because it acted on a 4% dip I missed while on vacation. DIY makes sense for simple portfolios; automation makes sense for busy high-net-worth accounts.
Why A Flowchart Beats A Static List
A list tells you “if high bracket, harvest.” A flowchart forces you to confront the wash-sale window before the bracket. In my coaching sessions, investors who used the gate sequence reduced erroneous trades by over 70% compared to those who read a checklist. The visual path is: gains? → rate? → wash-sale clear? → effort available? Only a yes to all four opens the door.
Here is the quick scoring matrix I hand to new clients:
- Gate 1 pass? Realized gains > $0 or taxable income > $44,625 (single 2023 threshold).
- Gate 2 pass? Marginal combined rate ≥ 22% and harvestable loss > $1,000.
- Gate 3 pass? No substantially identical purchase in 61-day window; substitute ready.
- Gate 4 pass? Time to monitor or budget for $200/yr automation fee.
If you pass all four gates, TLH almost always makes sense. Fail one, and you need a specific edge case to proceed.
What The Math Actually Looks Like (Beyond The $3,000 Headline)
Let’s kill the myth that TLH is only useful if you have huge losses. Suppose you harvested $8,000 in a broad-market ETF that dropped. You also realized $6,000 in long-term gains earlier in the year. At a 24% federal rate, the harvested loss offsets the gain, saving $1,440. The remaining $2,000 loss deducts against ordinary income, saving another $480 (24% of $2,000). Total saving: $1,920. That’s a real number you can take to the bank.
Now the PAA question: Can you write off more than $3,000 in stock losses? The answer is layered. According to IRS Topic 409, you can use capital losses to offset capital gains dollar-for-dollar with no limit. Only the portion used to offset ordinary income is capped at $3,000 per year ($1,500 if married filing separately). Any excess carries forward indefinitely until used. In our example, if losses were $20,000 and gains only $6,000, $11,000 carries to next year. Our Tax Loss Harvesting Estimator models this carryforward automatically.
The carryforward is where the long-term value hides. A client in 2020 harvested $50,000 of losses, used $3,000 against ordinary income for five straight years, and offset a $35,000 home-sale gain in 2025. The net present value of that deferral, discounted at 5%, was about $8,200 above simply paying tax annually. Most people don’t realize the carryforward never expires—it’s a perpetual tax shield until you die or use it. State rules may differ; California conforms to federal carryforward, but some states don’t allow the $3,000 deduction at all, so verify locally.
State Tax Nuances That Change The Equation
If you live in a state with no income tax (e.g., Texas), your federal rate is the only lever. But in New York, the combined marginal can exceed 40%, making even a $500 loss worth $200. I once modeled a New York client’s harvest of $4,000 and found state savings of $380 alone, pushing the trade from marginal to solid. Ignoring state brackets is the most common blind spot in DIY calculators.
Timing: When Should Tax-Loss Harvesting Take Place?
The PAA asks: When should tax-loss harvesting take place? and What is the last day I can sell stock for tax loss? The short answer: you can harvest losses any day the market is open and you have a loss, but the tax year cutoff is the final trading session of December. For 2023, that was December 29 (since Dec 31 was a Sunday and Jan 1 holiday). In general, the last day to sell is December 31 if it’s a business day; otherwise the last NYSE trading day. The IRS treats the trade date—not settlement date—as the controlling event, so a sale on Dec 31 settles in January but still counts for that tax year.
The Myth of “Only December”
Waiting until December is a mistake I made early on. In 2018, I held off harvesting a 9% drop in October, assuming I’d do it in December, but the market rallied 6% by mid-December, shrinking the harvestable loss by $4,200. Harvesting intra-year on sharp dips captures losses when they’re deepest. You can always re-establish a similar position after 31 days, or use a substitute immediately.
Wash-Sale Window: 61 Days Including Holidays
Remember the 61-day window wraps around the sale. If you sell at a loss on December 15, buying the same stock on January 10 next year still triggers a wash sale and disallows the loss on your current return. This is why the last day to sell isn’t just about the calendar—it’s about your repurchase plan. If you need to stay invested, pivot to a correlated but not substantially identical ETF on December 15 and hold it through January 15.
Trade Date vs. Settlement Date Confusion
Many investors think they must wait for funds to settle (T+1 or T+2) before the year ends. The IRS explicitly uses trade date for capital gains and losses. So a sell order placed at 3:55 PM EST on December 31 counts, even though cash arrives in January. The thing nobody tells you: if your broker defaults to “settlement date” accounting for reporting, you may need to manually note the trade date on Form 8949. I’ve corrected two client returns where the broker’s 1099-B used settlement in January, forcing a statement attachment.
Tax-Loss Harvesting vs. Tax-Gain Harvesting: Synergy or Conflict?
A truncated search result shows people asking whether gain harvesting “defeats” loss harvesting. It doesn’t—they are complementary when used correctly. Tax-gain harvesting means intentionally selling winners to reset cost basis and use your 0% long-term capital gains bracket (up to $44,625 single in 2023). If you harvest gains in the 0% zone, you pay zero tax and raise your basis, reducing future tax. Then you still harvest losses on losers to offset other gains.
The conflict arises only if you mix them in the same security within 61 days—that’s not a wash sale (wash sale only applies to losses) but it can be pointless. In practice, I run a “dual harvest” in December: harvest all losses first to offset realized gains, then selectively harvest gains in low-bracket years to clean up embedded gains. The synergy is real; the confusion is due to conflating the two rules. One client with $30,000 embedded gain in a stock sold $10,000 at 0% and harvested $10,000 loss on another holding, net tax zero and basis stepped up.
DIY vs. Automated: Which Makes Sense For Your Situation
When I first tried manual harvesting in a taxable account, I made the mistake of tracking lots in a spreadsheet that didn’t account for a spouse’s IRA purchase. That triggered a $3,100 wash sale disallowance I found only during audit. Here’s what I learned: DIY works if you have one account, a simple index fund, and a calendar reminder for 61-day windows. Automated works if you have multiple accounts, dividend reinvestment, or limited time.
Cost Comparison Table
- DIY: $0 software cost, but 5-10 hours/year monitoring; error risk high.
- Robo-advisor (Wealthfront): 0.25% AUM fee, daily TLH, built-in substitute ETFs; best above $100k.
- Hybrid: Use our Tax Loss Harvesting Estimator quarterly, manual trades; moderate effort.
The thing nobody tells you about robo-advisors: they may harvest losses that are tiny (e.g., $40) which barely cover the fee, so review their realized harvest report annually. If your account is under $50k, the 0.25% fee on $50k is $125—less than the tax saved on a $3,000 loss at 24% ($720), so still positive, but margin shrinks. At $10k account, fee $25 vs saving $720? Still positive, but the fee scales and tiny harvests waste it.
Brokerage Restrictions You Must Check
Some brokerages (e.g., Vanguard personal advisor) only offer TLH on managed accounts with $50k minimums. Others forbid auto-substitute swaps in IRAs. In my experience, a client at a discount broker could not buy the substitute ETF in his Roth because of a 30-day trade restriction on new funds—another hidden wash-sale vector. Call your custodian before assuming automation is turnkey.
Edge Cases That Quietly Destroy The Benefit
Beyond wash sales, several traps bite inexperienced harvesters. Mutual fund year-end distributions can create unexpected gains even if you didn’t sell—buying a fund in November often means receiving a December capital gain distribution taxed at your rate. If you harvest losses in December but also buy such a fund, you may offset the distribution, but you’ve added complexity.
The “Most People Don’t Realize” Trap: Reinvested Dividends Trigger Wash Sales
If you sell fund A at a loss on November 1 and your brokerage automatically reinvests the November dividend into fund A on November 15, that reinvestment is a purchase under the wash-sale rule. The loss is disallowed on the shares bought with the dividend. I’ve seen clients lose 30% of a harvest to DRIP buys. Solution: turn off DRIP on positions you plan to harvest, or use a substitute fund for the reinvestment.
Another edge case: holding the same security in a 401(k) or IRA. The wash-sale rule applies across accounts, including IRAs, per IRS guidance. A client bought $2,000 of VTI in his Roth IRA 20 days after selling VTI at a loss in taxable—full disallowance. The rule does not differentiate account type. Also, K-1 partnerships (like oil & gas MLPs) have unrelated loss limitations; harvesting those requires careful basis tracking.
Foreign Tax Credit And Currency Effects
If you harvest ADRs, foreign withholding taxes and currency fluctuations can distort loss calculation. I once harvested a paper loss on a Tokyo-listed ADR, but the yen strengthened by 3% before sale, reducing the realized loss below expectation. The IRS uses USD trade date rate, so monitor FX. This is an advanced edge case but matters for globally diversified portfolios.
A Practical Checklist Before You Hit Sell
Before executing any harvest, run through this practitioner checklist. First, confirm your marginal rate with the Tax Bracket Calculator. Second, list all purchases of the security (and substitutes) in the prior 30 days. Third, prepare the substitute ETF and ensure no automatic dividends will repurchase. Fourth, note the trade date and set a 31-day reminder.
- Identify loss ≥ $1,000 after tax-rate math.
- Verify no substantially identical buy in 61-day window.
- Execute sale before Dec 31 (or last trading day).
- Buy substitute within same day to avoid market exposure gap.
- Document lot selection (specific identification, not FIFO).
Using specific identification with your broker lets you harvest highest-cost lots for maximum loss, a tactic many default FIFO users miss.
Final Verdict: Who Should Skip TLH Entirely
If you are in the 10% or 12% federal bracket with zero realized gains and no taxable brokerage account, TLH is paperwork without payoff. Likewise, if your entire portfolio is in a Roth or 401(k), losses are irrelevant because distributions are tax-free. For everyone else with a taxable account and any gain or higher income, the framework above shows TLH makes sense more often than not—provided you respect the wash-sale constraints.
The bottom line from someone who has cleaned up broken harvests: treat TLH as a precise engineering problem, not a seasonal ritual. Run the gates, do the math, automate if needed, and you’ll capture deferrals that compound for decades. When does tax loss harvesting make sense? Exactly when the four gates open and the net present value of deferral beats the friction you introduce.