Tax errors come in two directions, and only one of them writes to you. Underreport your crypto gains and, sooner or later, a notice arrives: the IRS received a form you didn't reconcile, and the difference is now your problem. Overreport them — pay tax on profit you never made — and nothing arrives. No agency mails a letter saying the gain looks too high. The report is filed, the money is gone, and the error survives precisely because it inconveniences no one but you.
That second direction is where phantom gains live: profits that exist in a tax report but never happened, almost always because software lost track of what you paid for a coin and substituted zero. This guide is about the bill. Not the mechanics of how basis breaks — we cover that elsewhere — but the three distinct channels through which a broken basis inflates the tax you actually pay, how large the inflation runs in the published data, and the checks that catch it while it's still a bookkeeping problem rather than a refund claim.
This is not tax advice. It's an explanation of how the numbers behave, with links to primary sources. For decisions about your own return, talk to a tax professional.
Channel 1: You pay tax on money that never existed
The direct channel. Your taxable gain on a sale is proceeds minus cost basis. When software can't find the acquisition — most often after an unpaired transfer between your own wallets — it books the basis as $0, and the entire sale becomes gain. Buy 1 ETH for $2,400, move it to a hardware wallet, sell it for $3,000, and a report with a broken transfer shows a $3,000 gain instead of $600. The $2,400 difference isn't income by any definition. It's taxed anyway if it's filed.
The size of this channel scales with two things almost every crypto holder has: transfers between their own wallets, and time between buying and selling. To measure it, we ran the same cost-basis engine twice over the most public wallet cluster on Ethereum — identical data, identical classifications, the only difference being whether transfers between the entity's own wallets carried basis or dropped it. The naive run reported $44,975,739 more in lifetime realized gains than the basis-aware run. Same coins, same sales, same prices; the entire difference is phantom. The method, wallet list, and per-flag tx-hashes are published at that link, so the number can be checked rather than believed.
Most portfolios are not that cluster. But the published estimate for ordinary taxpayers is not small either: an analysis by the crypto tax firm Summ, covering 30,000 US users, estimated that broken basis chains left the average affected investor with $14,500 in overstated capital gains — overstated gains, to be clear, not overpaid tax; the tax cost is whatever your rate makes of them.
Channel 2: Long-term coins get taxed as short-term
The quieter channel, and often the more expensive one per dollar of gain. When an acquisition record is lost, the acquisition date is lost with it. Coins you held for years can surface in the report as acquired the day they landed in the new wallet — and a sale that qualified for long-term capital gains treatment gets computed at short-term rates instead, which for most filers are meaningfully higher.
This one doesn't require the basis dollars to be wrong at all. A transfer can be repaired well enough to restore the price you paid and still carry the wrong date. In one documented forum case, a user who moved BTC through a hardware wallet before selling on an exchange hit exactly this: the broken link both zeroed the basis and put multi-year holdings at risk of short-term treatment. When you audit your own report, check both columns — what you paid, and when.
Channel 3: Real losses vanish from the report
The channel nobody looks for, because everyone audits gains. A report that overstates your gains can simultaneously be hiding losses you actually realized — and a hidden loss is overpayment twice: once as the phantom gain you're taxed on, once as the real loss you never used to offset anything.
The mechanism is the same zero. If a coin's true basis was above its sale price, the sale was a loss; wipe the basis to $0 and the same sale reports as a gain. In our public-cluster experiment, the largest single distortions were sign flips — disposals where the basis-aware run showed a loss and the naive run showed a multi-million-dollar gain. The same flip happens at ordinary scale every time a coin bought near a top is moved between wallets and later sold below cost.
Why the overpayment goes unnoticed
Three reasons, each mundane.
The default feels safe. A $0 basis is the conservative assumption — conservative in the government's favor. Nothing about it looks reckless, no warning says "you are about to overpay," and dismissing the flags produces a clean-looking report. Clean is not correct.
The comparison never happens. In Coinbase and CoinTracker's 2026 Crypto Tax Readiness Report, a survey of 3,000 US crypto users, 76% said they knew their cost basis might need adjusting — but only 35% had ever adjusted it. The gap between suspecting and checking is exactly where the three channels operate.
The tools disagree with each other, so no single answer looks wrong. When a college professor ran identical transaction data through five crypto tax platforms, the results ranged from a $1,516 loss to a $2,696 gain against a true answer of a $1,266 loss. If you only ever see one tool's number, you have no reason to doubt it.
And the error does not stay fixable forever. A phantom gain discovered after filing generally means an amended return, and refund claims on Form 1040-X are subject to time limits — per the IRS, generally within 3 years of filing the original return or 2 years of paying the tax, whichever is later. Overpaid tax from a report nobody rechecked can simply age out of recoverability. The details of any amendment belong with a tax professional; the point here is only that the clock exists.
Three checks before you file
Each channel leaves a visible trace, and none of the checks require special tooling to start.
Pull the zeros, not just the warnings. Filter your report for disposals with a $0 or implausibly low basis, and separately sort all disposals by proceeds and inspect the top ten. Some zero-basis sales never raise a flag. Every large sale should show a basis you recognize and an acquisition date that matches when you actually bought — not when you last moved the coins.
Compare the report's total against your own memory of the year. If the software says you realized far more profit than you believe you took — or shows gains in a year you mostly sold at a loss — don't assume it knows better. It knows exactly what survived the import, nothing more.
Check the chain's version of events. Your wallets' actual history is the tiebreaker between what you remember and what the report claims. Run my free scan to see every transfer between your own Ethereum addresses that a tax tool would book as a zero-basis sale — read-only, no sign-up, with a tx-hash on every line.
If the checks turn something up, the repair path is documented: the tool-agnostic repair checklist for fixing the data, and the pre-filing remediation guide for carrying corrected numbers through to Form 8949.
FAQ
How do I know if I'm overpaying taxes on crypto?
Look for the three traces: disposals with a $0 or implausibly low cost basis, acquisition dates that match transfers rather than purchases (which convert long-term gains to short-term), and a reported total gain that exceeds what you believe you actually realized. Any of the three suggests the report is measuring sales against missing acquisition records rather than against what you paid, and the difference flows straight into the tax computed.
Can I get money back if I overpaid crypto taxes in a past year?
Potentially, through an amended return, but there are deadlines. The IRS states that refund claims on Form 1040-X must generally be filed within 3 years after the original return was filed or 2 years after the tax was paid, whichever is later. Whether amending is worthwhile in your case — and how to document a corrected basis — is a question for a tax professional, ideally with the broken transfers identified before that conversation.
Isn't a $0 cost basis safe, since it favors the IRS?
It's conservative, not correct, and it's your money making up the difference. Filing a knowingly inflated gain has no upside: the basis you actually paid is yours to substantiate with records, and per-wallet accounting since 2025 makes documented basis more important, not less. There's also a second-order cost — a zero basis usually comes with a wrong acquisition date, which can misclassify long-term holdings as short-term. Conservative on the dollars can still be wrong on the rate.
This is not tax advice, and none of the above is a substitute for a qualified tax professional who can look at your actual records.
If you'd rather know before filing whether your report is inflated, Verilot Check reads your Ethereum wallets directly on-chain and flags every transfer between your own addresses that would otherwise be booked as a sale, with a tx-hash on every line. Free and read-only — no wallet connection, no sign-up, up to 5 wallets and 500 events per scan, on Ethereum mainnet, Base or Arbitrum One (one chain per scan). Run my free scan.