Robo-advisor wash sales cost me a tax deduction

The dirty version starts when the same investor owns another account.

A dividend reinvestment at a second brokerage. A recurring purchase in an IRA. A spouse buying the same ETF. A manual trade made without checking what the robo-advisor is doing. Any of these transactions can create a wash sale and disallow the loss the algorithm just harvested.

This is the central problem with robo advisor tax loss harvesting wash sale strategies: the software may control one account, while the IRS looks at the broader transaction pattern.

The loss is not usually destroyed permanently. It is deferred. But deferred tax benefits are not the same as current tax benefits, and the difference matters when the entire strategy was built around reducing this year’s taxable income.

The tax-loss harvesting promise is real. So is the blind spot.

Tax-loss harvesting is straightforward in principle.

If an investment has fallen below your purchase price, a robo-advisor can sell it and realize the capital loss. It then buys a replacement security designed to preserve similar market exposure. The realized loss can offset capital gains. If your net capital losses exceed your gains, up to $3,000 can generally be deducted against ordinary income in a tax year, with unused losses carried forward.

That is the tax mechanism. The investment mechanism is portfolio maintenance.

You do not want to sell a broad-market position, sit in cash, and hope to remember to reinvest later. The robo-advisor sells the declining security and moves the proceeds into a proxy fund or another ETF in the same asset class. Your allocation remains close to target. Your tax lot changes.

In a single managed account, this can be efficient. The platform knows which lots it sold and which replacement security it bought. It can avoid repurchasing the same or a substantially identical investment inside that account.

The weakness is obvious once you stop viewing the robo-advisor as a complete picture of your finances.

Most robo-advisors do not have a reliable view of every taxable brokerage account, IRA, bank sweep, dividend reinvestment setting, or spouse account connected to your household. The algorithm can coordinate its own trades. It generally cannot coordinate transactions it cannot see.

Automated tax-loss harvesting is only as clean as the transaction data behind it. One unmanaged account can turn a precise algorithm into a partial tax record.

The term “wash sale” refers to a loss-generating sale followed by a purchase of the same or a substantially identical security within the relevant 61-day window. That window covers the 30 days before the sale, the day of the sale, and the 30 days after it.

The rule also works in reverse. A purchase made before the robo-advisor sells the losing position can matter. You do not get a free pass because the automated sale happened later.

How a robo-advisor wash sale happens

Consider a simplified example.

Your robo-advisor owns an ETF that has declined. It sells that ETF at a loss and buys a replacement ETF to maintain your target allocation. That transaction may be designed to avoid a wash sale within the managed account.

But you also hold the original ETF in a separate brokerage account. A dividend from that account is set to reinvest automatically. The reinvestment buys a small amount of the same ETF during the 30 days after the robo-advisor’s sale.

The loss can be disallowed to the extent of the replacement purchase.

The amount is not necessarily lost forever. Instead, the disallowed loss is generally added to the cost basis of the replacement security. The tax benefit moves forward until that new position is sold in a transaction that does not create another wash sale.

That sounds harmless. It is not always harmless.

A basis adjustment can be useful eventually, but it does not deliver the deduction when you expected it. It can also create additional recordkeeping complexity, particularly if the replacement purchase is small, occurs in another account, or involves a position you intend to hold for years.

The practical issue is timing. Tax-loss harvesting works partly because losses may be valuable now. A deduction postponed into an uncertain future has a lower practical value than a deduction you can use today.

The most common automated tax-loss harvesting mistakes

The mistakes are usually operational, not mathematical:

1. Leaving dividend reinvestment turned on elsewhere.

A small automatic purchase can be enough to create a wash sale. The dollar amount of the reinvestment does not make the transaction irrelevant.

2. Continuing a recurring investment plan.

Weekly or monthly purchases in another brokerage account can overlap with the robo-advisor’s harvest. You may not know the sale occurred until the tax documents arrive.

3. Buying the same ETF manually.

Investors sometimes replace a fund sold by the robo-advisor because they believe the decline makes it attractive. That can defeat the tax result.

4. Ignoring IRA and spouse accounts.

The relevant transaction may not occur in the taxable account where the loss was generated. External retirement accounts and a spouse’s account can still create coordination problems.

5. Assuming the platform sees everything.

Account aggregation is not the same as tax-lot control. A dashboard may display another account without being able to prevent every trade inside it.

6. Treating similar funds as automatically safe.

A replacement security can be different without being clearly outside the “substantially identical” question. The precise treatment of overlapping ETF exposures is not always mechanically obvious.

That last point is where marketing language becomes less useful. A robo-advisor may advertise tax-loss harvesting as an automated feature. It cannot turn an unsettled tax classification into a guaranteed outcome.

The 61-day rule is simple to state and easy to violate

The basic robo advisor wash sale rules are not complicated:

• The security is sold at a loss.
• The investor buys the same or a substantially identical security within 30 days before or after that sale.
• The loss is disallowed for the current deduction.
• The disallowed amount is generally added to the basis of the replacement security.

The difficulty is that investors think in account boundaries. The tax rule does not necessarily respect the boundaries you use in your app.

You may have:

• A taxable robo-advisor account.
• A self-directed brokerage account.
• An IRA with automatic contributions.
• An employer stock plan.
• A spouse’s brokerage account.
• A portfolio tracking app that encourages manual rebalancing.
• Dividend reinvestment enabled across several platforms.

The more places you trade, the more fragile automated harvesting becomes.

This does not mean you should close every account or refuse to use a robo-advisor. It means you need to decide whether the convenience of automation is compatible with your transaction habits.

If you hold one taxable portfolio and rarely trade, the setup is relatively clean. If you run several brokerage accounts, actively trade ETFs, and keep reinvestment schedules running, the yield drag from a failed harvest may be self-inflicted.

A simple timeline

Suppose the robo-advisor sells a losing fund on Day 0.

The calendar is not the hard part. The hard part is knowing what every account is doing during the calendar.

Why the tax benefit can justify automation—but not blind trust

Systematic tax-loss harvesting can generate meaningful tax alpha for some investors. Financial research cited in the underlying material estimates average annual tax alpha in the rough range of 0.72% to 1.08% under certain conditions.

That is not a universal return. It is not a coupon. It is not available in the same way to every investor.

The benefit depends on several variables:

• Whether you have realized capital gains to offset.
• Whether your tax bracket makes the losses valuable.
• Whether the harvested losses can be used against ordinary income.
• Whether you continue contributing new money.
• Whether markets provide enough declining positions to harvest.
• Whether you avoid wash sales.
• Whether the platform’s fees and portfolio construction create additional drag.
• Whether the assets are held in a taxable account rather than an IRA.

This is where we need to separate a feature from an outcome.

A robo-advisor can execute the trades. It cannot create tax value where no taxable income exists to offset. It cannot make a wash sale impossible across accounts it does not control. It cannot guarantee that a replacement ETF will produce the exact exposure you want under every market condition.

The upside is asymmetric for an investor with taxable gains, long holding periods, and disciplined account coordination. The upside is much smaller for someone with no gains, low taxable income, or a portfolio dominated by tax-advantaged accounts.

If most of your investing happens inside an IRA, automated tax-loss harvesting is not the main reason to choose a robo-advisor. The tax-loss feature applies to taxable investing. Paying attention to it while ignoring fees, asset allocation, cash management, and withdrawal rules is backwards.

How to prevent a wash sale without dismantling your portfolio

You do not need a complicated tax command center. You need a controlled process.

First, map every account that can trade

Start with the household, not the robo-advisor dashboard.

List every account that may buy securities, including accounts that do not feel like active investment accounts. Retirement accounts count. Spouse accounts count. Employer plans may matter depending on the securities involved and the specific transaction. Dividend reinvestments count because they are purchases.

Then identify which accounts hold the same funds or securities as the robo-advisor.

The target is not a perfect net-worth spreadsheet. The target is a transaction map.

Second, decide who controls the exposure

There are two workable approaches.

Approach one: centralize the exposure.

Let the robo-advisor manage the taxable allocation and stop buying the same securities elsewhere. This reduces the number of moving parts.

Approach two: separate the securities by account.

Use different funds or asset exposures in different accounts, with care around substantially identical positions. This can work, but it requires more judgment and more monitoring.

The dangerous approach is accidental duplication. That is when you let the robo-advisor harvest automatically while continuing to make uncoordinated purchases elsewhere.

Third, disable automatic reinvestment where necessary

Dividend reinvestment is useful. It is also easy to forget.

If you are trying to preserve a harvested loss, automatic reinvestment in an external account can create the exact purchase you were trying to avoid. You may choose to receive dividends in cash temporarily, or redirect them into a security that does not create the same overlap. The correct choice depends on your portfolio and tax situation.

The point is control. A reinvestment setting should be a deliberate portfolio decision, not background software.

Fourth, keep a transaction log

Your brokerage statements and year-end tax forms are not a substitute for an operating record.

For each harvested position, record:

• The security sold.
• The sale date.
• The number of shares sold.
• The loss amount.
• The replacement security.
• The date of the replacement purchase.
• Any recurring purchases or reinvestments scheduled elsewhere.
• Whether the replacement security remains in the portfolio.

This is particularly useful when you use portfolio tracking software. The software can help you see holdings, cost basis, allocation, and account activity in one place. It should not be treated as a tax-law oracle.

Portfolio tracking tax loss harvesting workflows are valuable when they expose conflicts before trades happen. They are less valuable when they merely display a wash sale after the fact.

Fifth, review the platform’s actual controls

Marketing pages often describe the feature at a high level. You need the mechanics.

Ask:

• Does the platform use proxy securities during harvesting?
• How does it define replacement assets?
• Can you opt out of specific securities?
• Can you turn off tax-loss harvesting without closing the account?
• Does the platform monitor linked external accounts, or merely display them?
• How are disallowed losses reported?
• Does the platform provide tax-lot details?
• What happens when you withdraw or transfer assets?

Do not confuse “automated” with “comprehensive.” Automation usually means the platform follows a ruleset within a defined system. It does not mean the platform assumes responsibility for every account in your household.

The cheapest way to avoid a wash sale is often not better software. It is fewer overlapping purchase instructions.

What happens after the loss is disallowed?

A wash sale does not usually erase the economic loss. It changes its tax timing.

Assume you sell a position for a $1,000 loss, but a replacement purchase causes the loss to be disallowed. The disallowed amount is generally added to the cost basis of the replacement security. When that replacement security is later sold, the adjusted basis can affect the eventual gain or loss.

This creates two distinctions that investors often blur:

1. The investment loss.

The asset declined in market value. That economic result already happened.

2. The tax recognition of the loss.

The tax code may delay when that loss can be used.

If the replacement security is eventually sold at a gain, the higher basis can reduce that gain. If it is sold at a loss, the adjusted basis can affect the size of the recognized loss. The timing and subsequent transactions determine the practical result.

The phrase “cost me a tax deduction” is therefore directionally right but technically incomplete. A wash sale may cost you the deduction this year. It may not cost you the loss permanently.

That distinction matters for accurate planning. It also matters for trust. Robo-advisor sales pages tend to emphasize the benefit. Investor forums sometimes overcorrect and claim the loss has vanished forever. Neither version is adequate.

Should you turn off automated tax-loss harvesting?

There is no universal answer. There is a decision rule.

Keep it enabled when all of the following are broadly true:

• You invest in a taxable account.
• You have gains or taxable income that make losses potentially useful.
• You hold a diversified portfolio with positions that can be harvested.
• You do not routinely buy the same securities in other accounts.
• You are willing to monitor reinvestment and recurring purchase settings.
• The platform’s fee and portfolio design do not create excessive drag.

Consider turning it off, or managing the process manually, when:

• You trade the same ETFs across multiple brokerages.
• You regularly contribute to an IRA containing overlapping investments.
• You and your spouse invest independently without coordination.
• You want complete control over tax lots.
• The harvesting feature creates replacement holdings you do not actually want.
• You lack capital gains and the expected tax value is limited.
• You are likely to keep buying through automated plans during the restricted window.

Manual management is not automatically superior. It simply moves the burden from the algorithm to you. If you do not maintain accurate records, manual harvesting can produce its own errors.

A robo-advisor is most useful when it removes repetitive work without hiding important decisions. If the automation encourages you to stop thinking about cross-account exposure, the convenience can become an opportunity cost.

The platform decision is really a behavior decision

Investors often compare robo-advisors as if the main question were which algorithm has the better tax-loss harvesting engine.

That is only half the analysis.

The better question is whether the platform fits your behavior.

If you want a managed taxable portfolio, rarely trade elsewhere, and can keep duplicate purchases out of the system, automated harvesting may be a sensible tool. It can maintain allocation, harvest losses, and reduce the need for manual lot selection.

If you want to trade ETFs in several accounts, keep automatic contributions active, and make independent portfolio decisions throughout the year, the same feature may produce friction. You will need to coordinate every overlapping purchase or accept that some harvested losses may be deferred.

This is not a reason to fear robo-advisors. It is a reason to reject feature worship.

Tax-loss harvesting is a process, not a performance guarantee. The relevant return is the after-tax outcome after fees, replacement trades, account coordination, and eventual disposition of the assets.

When evaluating a robo-advisor, I would rank the issues in this order:

1. Portfolio construction.

The allocation must make sense before tax optimization enters the conversation.

2. Total cost.

Advisory fees, fund expense ratios, cash allocations, and trading frictions all contribute to yield drag.

3. Taxable-account functionality.

Look for tax-lot control, harvesting logic, proxy securities, and clear reporting.

4. Account coordination.

A sophisticated algorithm is less useful if your other accounts continue buying the same securities.

5. Operational transparency.

You should be able to understand what was sold, what replaced it, and how the platform reports the result.

6. Your ability to follow the system.

A theoretically superior process fails if it conflicts with your habits.

The strict conclusion

The robo-advisor did not necessarily fail when a wash sale disallowed your harvested loss. The system may have executed its own trades correctly. The failure may have been treating one managed account as if it represented the entire household.

That is the mistake to eliminate.

If you use automated tax-loss harvesting, control the 61-day window across every account you influence. Disable or redirect overlapping reinvestments. Stop recurring purchases that conflict with harvested positions. Track replacement securities and cost basis. Do not assume account aggregation equals tax coordination.

Then make the binary choice:

Either let the robo-advisor control the relevant exposure and accept its rules, or keep trading the same securities elsewhere and manage the tax consequences yourself.

The middle ground—automated harvesting combined with untracked duplicate purchases—is where the tax deduction goes to die.