Who Actually Reports Crypto to the IRS — and What the Data Reveals

One thing that's been on my mind since writing about how taxes actually work is this: the whole system depends on people reporting honestly. For traditional income — your W-2 salary, your bank interest — that's mostly automated. Your employer and your bank send data directly to the IRS before you even file.

But cryptocurrency? That's been a different story.

A recent study using actual IRS population-level data found that only 32–56% of cryptocurrency holders may be reporting their gains to the IRS. That's a massive gap — and the reasons behind it are more interesting than I expected.

The study is by Jeffrey Hoopes, Tyler Menzer, and Jaron Wilde (2026), titled "Who reports cryptocurrency to the IRS?" It's one of the first papers to use U.S. administrative tax data at a population level to study crypto investors.

Here's the takeaway that stuck with me: the crypto reporting gap isn't really a story about people cheating on their taxes. It's a story about what happens when a new asset class shows up and the plumbing — the automatic forms, the broker reporting, the checkbox prompts — hasn't caught up yet. There's striking evidence for this from outside crypto: a Danish audit experiment (Kleven et al., 2011) found tax evasion was 0.3% on third-party-reported income but 37% on self-reported income — same people, same country, same year. Once you see the gap that way, almost every weird finding in this crypto paper starts to make sense.

What makes the timing interesting: crypto sits at a policy crossroads. President Biden signed an executive order in 2022 calling for a "whole-of-government approach" to digital assets. Meanwhile, President Trump has floated the idea of a strategic Bitcoin reserve and even eliminating taxes on certain crypto gains. Whether the future brings tighter reporting or looser taxation — understanding who's currently in the system (and who isn't) matters for either direction.

Crypto is taxed as property — and that matters more than you think

If you remember from the tax basics post, the IRS cares about all your income — earned and unearned. Capital gains from selling stocks, for example, get reported on Form 8949 and show up on your 1099-B.

Cryptocurrency follows the same logic, but with a twist. The IRS issued Notice 2014-21, classifying crypto as property — not currency, not a security. That classification means every time you sell Bitcoin for a profit, it's a capital gain , just like selling a stock. You report it on the same Form 8949.

But here's the thing — while your broker automatically sends 1099-B forms for stock trades, crypto exchanges haven't been required to do the same for most of the period this study covers (2013–2021). That means crypto reporting has been largely voluntary.

And voluntary reporting is where things get interesting.

The reporting gap is real — and it's huge

The researchers compared IRS data against external estimates of crypto ownership from surveys, academic studies, and Coinbase's public filings. The gap is striking.

Between 44–68% of crypto holders may not be reporting their gains at all. And the population of Reporting Crypto Sellers (as the researchers define them) is growing rapidly — but it's still a fraction of the total.

This isn't surprising when you think about it. Without automatic third-party reporting, the system relies on taxpayers knowing they owe taxes on crypto, understanding how to report it, and choosing to do so. That's a lot of friction for an asset that many people bought through an app on their phone.

Who are these crypto sellers? Younger, lower-income, and less financially sophisticated

Here's where the data gets really interesting. The researchers compared Reporting Crypto Sellers to Reporting Non-Crypto Investors (people who sell traditional stocks and bonds) across demographic attributes.

On average, crypto sellers report lower income, are younger, are more likely to be students, and are less likely to itemize their tax deductions. All of these attributes are generally associated with less financial sophistication.

That clicked for me because it connects back to something from the investing post: the kind of investor you are shapes everything about how you approach the market. Crypto has attracted a different demographic than traditional equity markets — and that has real implications for tax policy.

The meme stock connection

One of the most interesting findings is that crypto sellers also trade significantly more meme stocks than other investors.

Remember the GameStop/WallStreetBets frenzy of 2021? The researchers found that between 2.5% and 5.8% more of crypto sellers' stock portfolios were meme stocks compared to other investors — and this held even after controlling for demographics. In fact, the top holding of the MEME ETF at the time was Riot Blockchain, a cryptocurrency mining company.

This pattern suggests that crypto and meme stocks attract similar investor profiles — people who are drawn to momentum, sentiment, and social-media-driven trading rather than fundamental analysis.

The researchers aren't saying crypto sellers cause meme stock trading or vice versa. But the overlap is consistent and significant — and it matters because it tells regulators something about who they're designing policy for.

The checkbox that changed everything

In 2019, the IRS introduced a virtual currency question on tax returns. At first it was buried on Schedule 1. But in 2020, it was moved to the very first item on Form 1040 — right after your name and Social Security number.

The result? A significant increase in cryptocurrency reporting after the checkbox was introduced. People who were self-preparing their taxes (no paid preparer) saw an even larger increase.

This is a concept called "smart return" design — the idea that how you design a form changes how people respond to it. Moving the crypto question to the front of the 1040 was a signal: we're watching.

That stayed with me because it shows how a tiny UX change — literally moving a yes/no question to the top of a form — can shift behavior at a population level. The IRS didn't need to audit more people. They just needed to make the question unavoidable.

The 52-week puzzle: crypto sellers don't time their taxes

Here's one of the more subtle findings. Under U.S. tax law, if you hold an asset for more than one year (52 weeks), your gains are taxed at the lower long-term capital gains rate. For traditional stocks, there's a well-documented phenomenon called "bunching" — investors cluster their sales right after the 52-week mark to qualify for the lower rate.

When the researchers looked at crypto sellers' traditional stock trades, they saw the same bunching pattern as everyone else. But when they looked at those same people's crypto trades — the pattern largely disappeared. Crypto sellers appear to be less sensitive to long-term capital gain holding period incentives for their crypto, even though they respond normally for their stock trades.

That's a puzzle. The same people are sophisticated enough to time their stock sales for tax benefits but don't do the same for crypto. One possible explanation: crypto is newer, more volatile, and people trade it more impulsively. Another: the tax rules for crypto simply aren't as well understood.

The difference between short-term and long-term rates can be enormous. If you're in the 24% bracket, holding Bitcoin for 366 days instead of 364 could save you 9 percentage points on your gains. That's real money — but many crypto sellers aren't taking advantage of it.

Try it yourself — plug in a crypto trade and see how much holding for over a year could save you:

The tax-loss harvesting loophole

Here's something that caught my attention: during the study period, wash sale rules did not apply to cryptocurrency.

For stocks, if you sell at a loss and buy back the same stock within 30 days, the IRS disallows the loss deduction — that's the wash sale rule. It prevents people from harvesting tax losses while keeping the same position.

Because crypto was classified as property (not a security), the wash sale rule didn't apply. This meant you could sell Bitcoin at a loss on Monday, buy it back on Tuesday, claim the tax deduction, and keep your position. Research by Cong et al. (2023) documented that some sophisticated investors were doing exactly this.

This is a genuine loophole — and it creates a two-track system. Sophisticated investors who understood the rules could harvest losses freely. Less sophisticated investors (the majority, based on the demographic data) didn't even know this was possible.

Note: New rules under the Infrastructure Investment and Jobs Act are closing this loophole. Starting January 2025, custodial crypto brokers (Coinbase, Kraken, Gemini, etc.) are required to report gross proceeds to the IRS on the new Form 1099-DA; cost basis reporting follows in January 2026. The original rule covering DeFi and non-custodial brokers was rolled back via the Congressional Review Act, so non-custodial platforms are out of scope for now. Wash-sale extension to digital assets has been proposed in various tax bills but isn't yet enacted federally as of 2026. The landscape is still moving fast.

The computer science angle: how do you find crypto in tax data?

This is where it gets fun for me as someone who also studies computer science.

The researchers faced a surprisingly technical problem: IRS data doesn't have a neat "cryptocurrency" checkbox on Forms 8949 and 1099-B for most of the study period. So how do you identify which transactions are crypto?

They used textual analysis — essentially, string matching on the descriptions that taxpayers and brokers write on these forms.

If you've studied algorithms, this is essentially the naive string matcher problem from discrete mathematics. You have a text (the transaction description) and a pattern (crypto-related keywords), and you need to determine if the pattern occurs in the text.

The naive string matcher checks every possible position in the text, making it O(nm) in the worst case — where n is the length of the text and m is the length of the pattern. For a small dictionary of crypto keywords against short form descriptions, this works fine. But the concept scales to much bigger problems.

What the researchers built is really a classification pipeline — a system that takes unstructured text data and assigns each record to a category (crypto or not crypto). In computer science terms, this is a binary classification problem:

The parallels go deeper. In the theory of computation, we study decidable problems — questions where an algorithm can always give a definitive yes/no answer. "Does this transaction description contain a crypto keyword?" is decidable. But "Did this taxpayer actually sell cryptocurrency?" is harder — it depends on information that may not be in the description at all. The researchers acknowledge this limitation: they can only observe what's reported, not what's hidden.

It's the same gap that shows up in computability theory. Some problems are decidable (you can always verify). Others require you to trust the input — and when the input is voluntary self-reporting, trust has limits.

That gap is the whole story of this paper, in miniature. You can detect what's written down — keywords on a form, descriptions on a 1099-B. You can't detect what nobody bothered to write down in the first place. That's why the fix isn't smarter algorithms or better audits — it's making sure the data shows up automatically, before anyone has the chance to leave it out. (If you've taken a theory of computation course, Sipser's textbook frames this as the line between decidable and recognizable problems — some questions you can always answer; others, only when the input cooperates.)

You can model the reporting decision as a tiny state machine. Drag the nodes around to see it:

The q0 → q0 self-loop is the unreported trade. The classifier never sees it. No string matching algorithm — naive, KMP, Boyer-Moore, anything — can recover what isn't there.

Third-party reporting is the real game-changer

The single most effective tool for tax compliance isn't audits or penalties — it's third-party reporting. As we saw at the top of this post, when a third party (your employer, your bank, your broker) reports your income directly to the IRS, compliance shoots up to ~95%. The IRS's own 2022 tax gap projections show the inverse at scale: a $696B annual gross gap, with $539B (77%) of it coming from underreporting on returns that did get filed.

For most of the study period, crypto existed in the self-reported zone. Exchanges like Coinbase sent some 1099-B forms, but it was inconsistent and voluntary. The researchers found that IRS-reported crypto volume was only a fraction of what the Bitcoin blockchain actually recorded.

The Infrastructure Investment and Jobs Act (signed 2021, with final Treasury regulations issued in 2024) changes this. Custodial crypto brokers — the major exchanges — are required to issue Form 1099-DA for gross proceeds starting January 2025, with cost basis reporting following in January 2026. This single change will likely close the reporting gap more than any checkbox, penalty, or educational campaign ever could — though Treasury rolled back the original rule for DeFi and non-custodial platforms via the Congressional Review Act, so the gap won't fully close.

But aren't we just describing young people?

I want to flag one honest pushback to all of this. A skeptical reader could look at the "younger, lower-income, less likely to itemize" finding and say: that's not a crypto problem, that's just a first-time-investor problem. Stocks would look the same if you ran this study in 1985.

That's fair. But Lusardi et al.'s research on financial literacy among the young shows that fewer than a third of young adults grasp basic interest, inflation, or diversification — and yet they still hit ~95% compliance on their W-2 wages. The system carries them. They're not "more sophisticated" with stocks; their broker is just doing their reporting for them. So the demographic finding isn't really about who's bad at taxes — it's about who falls through when the plumbing isn't there. Same story, told from a different angle.

What this means for the average crypto holder

If you own crypto — or are thinking about it — here are the practical takeaways from this research:

1. Every crypto sale is a taxable event. Selling Bitcoin, swapping ETH for another token, even paying for coffee with crypto — all of these trigger capital gains or losses that need to be reported on Form 8949.

2. Holding period matters. If you hold for more than one year, you get the lower long-term capital gains rate. The data shows most crypto sellers aren't taking advantage of this — don't be one of them.

3. Third-party reporting is coming (mostly). Starting January 2025, custodial exchanges (Coinbase, Kraken, Gemini, etc.) report gross proceeds to the IRS on Form 1099-DA — cost basis follows in 2026. The DeFi/non-custodial piece was rolled back via the Congressional Review Act, so on-chain swaps and self-hosted wallets aren't covered. If you've been underreporting on a major exchange, the gap between what you filed and what the IRS knows is about to close.

4. Track your cost basis. Unlike stocks where your broker tracks everything, many crypto transactions (especially from earlier years, DeFi, or wallet-to-wallet transfers) have messy cost basis records. Start organizing now.

A few things I'm taking away

• Only 32–56% of crypto holders may be reporting their gains to the IRS — the gap between ownership and reporting is enormous and it's driven by both ignorance and the lack of automatic reporting
• Crypto is taxed as property under IRS Notice 2014-21, which means every sale triggers a capital gain or loss reported on Form 8949 — the same form you'd use for stocks
• The average crypto seller is younger, earns less, and is less likely to use a tax preparer or itemize deductions compared to traditional equity investors — this demographic profile has real implications for how policy should be designed
• Crypto sellers also trade significantly more meme stocks, even after controlling for demographics — suggesting these asset classes attract similar momentum-driven, social-media-influenced investor profiles
• The IRS virtual currency checkbox, moved to the front of Form 1040 in 2020, significantly increased crypto reporting — proving that form design is a powerful behavioral nudge
• Crypto sellers time their stock sales around the 52-week holding period for tax benefits, but don't do the same for their crypto — suggesting the tax rules for digital assets aren't well understood
• During the study period, wash sale rules didn't apply to crypto, creating a legal tax-loss harvesting loophole that sophisticated investors exploited while less sophisticated ones didn't know it existed
• The researchers identified crypto transactions using textual pattern matching on Form 8949 descriptions — the same string matching algorithms you'd study in a computer science course, applied to a tax compliance problem
• Third-party reporting is the most effective compliance tool — when brokers report directly to the IRS, compliance jumps to ~95% (Kleven et al., 2011 — same people, just different plumbing) — and the Infrastructure Investment and Jobs Act is bringing this to custodial crypto exchanges starting January 2025 via Form 1099-DA, though the DeFi/non-custodial piece was rolled back via the Congressional Review Act
• The classification challenge in this research mirrors problems in CS: decidable questions (does this text contain "BTC"?) are easy, but the real question (did this person actually trade crypto?) depends on information that may not be observable
• Government interventions always come with trade-offs — the checkbox nudge worked but also revealed that many people who checked "yes" for crypto didn't actually report any transactions, highlighting the continued need for education

That last point about third-party reporting is the one that reframed things for me. We spend so much time debating penalties, audits, and education — but the research consistently shows that the simplest fix is just making the information flow automatic. When your broker tells the IRS what you sold, you have every incentive to report it accurately. It's the same principle behind W-2 withholding: the system works best when it doesn't rely on memory, knowledge, or goodwill alone.

The question I'm sitting with, though, is what crypto trades away in exchange. A lot of people came to crypto specifically because it didn't route through a broker that reports to a government — that was the whole point of Eric Hughes's A Cypherpunk's Manifesto back in 1993: "privacy is necessary for an open society in the electronic age." Mandatory 1099-DAs solve the compliance gap, but they also quietly retire one of the things crypto was originally sold on. That trade isn't hypothetical either — Coin Center's §6050I challenge is back in district court after a Sixth Circuit win, arguing that mandatory crypto transaction reporting violates the Fourth Amendment. Whether the trade is fair depends on what you wanted from crypto in the first place.

Sources

• Hoopes, J.L., Menzer, T.S., & Wilde, J.H. (2026). Who reports cryptocurrency to the IRS? Review of Accounting Studies.
• Cong, L.W., Landsman, W., Maydew, E., & Rabetti, D. (2023). Tax-loss harvesting with cryptocurrencies. Journal of Accounting and Economics, 76(2–3).
• Dowd, T., & McClelland, R. (2019). The bunching of capital gains realizations. National Tax Journal, 72(2), 323–358.
• Kleven, H.J., Knudsen, M.B., Kreiner, C.T., Pedersen, S., & Saez, E. (2011). Unwilling or unable to cheat? Evidence from a tax audit experiment in Denmark. Econometrica, 79(3), 651–692.
• Lusardi, A., Mitchell, O.S., & Curto, V. (2010). Financial literacy among the young. NBER Working Paper No. 15352.
• Sipser, M. (2012). Introduction to the Theory of Computation, 3rd ed. — Ch. 4 (Decidability).
• IRS (2024). Final regulations and IRS guidance for reporting by brokers on sales and exchanges of digital assets.
• IRS (2025). About Form 1099-DA, Digital Asset Proceeds From Broker Transactions.
• IRS (2024). 2022 Tax Gap Projections.
• Hughes, E. (1993). A Cypherpunk's Manifesto.
• Coin Center. Constitutional challenge to §6050I crypto reporting.
• IRS Notice 2014-21 — Virtual Currency Guidance.
• Infrastructure Investment and Jobs Act (2021), Pub. L. No. 117-58, §§ 80603.