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. And the picture it paints is nuanced — it's not just about people dodging taxes. It's about a new asset class colliding with an old system, and what happens when the rules haven't fully caught up.

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 in 2025, crypto brokers are required to report transactions to the IRS via 1099 forms, and wash sale rules are being extended to digital assets. The landscape is changing 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.

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. Research consistently shows that when a third party (your employer, your bank, your broker) reports your income directly to the IRS, compliance rates shoot up to 95%+.

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, phasing in through 2025-2026) changes this. Crypto brokers will be required to send 1099 forms to the IRS, just like stock brokers do. This single change will likely close the reporting gap more than any checkbox, penalty, or educational campaign ever could.

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. Starting in 2025-2026, your crypto exchange will report your transactions directly to the IRS. If you've been underreporting, 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%+ — and the Infrastructure Investment and Jobs Act is bringing this to crypto starting in 2025
• 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.

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.
• IRS Notice 2014-21 — Virtual Currency Guidance.
• Infrastructure Investment and Jobs Act (2021), Pub. L. No. 117-58, §§ 80603.