On September 16, one committee of one chamber of one legislature placed a date on its calendar. That is the entirety of what can be verified.
The House Ways and Means Committee will hold a markup on crypto tax rules. No bill number. No legislative text. No named sponsor. No party breakdown. No Joint Committee on Taxation score. No effective date. No original source document. One procedural fact, and three sweeping conclusions attached to it: the rules could affect federal revenue, could affect regulatory clarity, could affect America's standing in global digital assets.
I have spent eleven years reading protocol code and statutory language side by side, and I can tell you which of the two is easier to audit. Code executes exactly what it says. A markup notice says nothing and executes later, if at all. What arrived this week was not a policy signal. It was a scheduling artifact wearing the vocabulary of inevitability.
That mismatch — one procedural fact producing three macro conclusions — is the only part of this story worth analyzing. Everything else is a forecast filmed as a report.
What a markup actually is
Start with the word. Markup is not a vote to pass. It is a committee's line-by-line rewrite session, the procedural starting gun, and the cheapest, least informative milestone in the entire legislative chain.
The full path runs: subcommittee drafting, full committee markup and vote, House floor vote, Senate referral to its tax-jurisdiction counterpart, Senate floor where ordinary legislation needs sixty votes to end debate, conference committee reconciliation of two different texts, presidential signature. Any stage can kill the bill. Most bills die at the third or fourth. Attendance at the first stage is not evidence about the sixth.
This matters because coverage of September 16 compressed that chain into a single image — "Congress is moving on crypto taxes." Congress is not moving. A committee is scheduling. Those are not synonyms, and the gap between them is where most retail capital gets destroyed.
Jurisdiction is the second omitted fact. Tax legislation originates in Ways and Means, but the Senate counterpart is the Finance Committee, not Banking, not Agriculture. Tax bills run on a different rail than securities bills. That rail has its own schedule, its own staff, and its own political logic, and none of it is visible in a September calendar entry.
There is a fast lane, too. Ordinary Senate bills need sixty votes. Reconciliation measures, passed under special budget procedures, need fifty-one. The large tax packages of the past two decades have mostly traveled the fast lane. If crypto provisions get attached to a reconciliation vehicle, their probability of becoming law rises sharply. If they move standalone, it collapses.
We do not know which vehicle this is. The notice does not say. That single unknown is worth more to your portfolio than any rate table published this month.
The information base is thin enough that the correct output is questions, not conclusions
Two data points exist. One is a fact: a markup is scheduled for September 16. One is an opinion: the rules could affect federal revenue, regulatory clarity, and national competitiveness. Nothing else survived — no clause text, no fiscal score, no sponsor, no year context beyond a date.
I want to be precise about what that means. A fiscal score is not a formality. The Joint Committee on Taxation's estimate is the only document that converts legislative language into a number, and that number determines whether a provision survives the process. A provision that raises revenue is a bargaining chip. A provision that loses revenue is a target. Without the score, "could affect federal revenue" is not a finding. It is a placeholder that can accommodate a rounding error or a structural shift of tens of billions, and the reader cannot tell which.
This is not a new failure mode. When I was an undergraduate I found a signature-malleability flaw in the v1 contracts of a major exchange protocol — improper nonce handling that permitted double-spend style attacks on relayed orders. I wrote a proof of concept, submitted it, and was told by the core team that I had probably misread the code. The flaw was real. It was patched in v2 after a delay that cost early users. The lesson I took from that episode was not about the developers. It was about evidence standards: I stopped arguing from summaries and started arguing from the artifact itself. A release note is not a patch. A roadmap is not a diff. A markup notice is not a statute.
So I will not tell you what this bill does. I will tell you the five mechanical questions that decide what it does, and how each one transmits into asset prices, protocol architecture, and the cost of holding anything on-chain.
Question one: is there a de minimis exemption?
This is not a rounding detail. It is the boundary between a currency and a collectible. Under current treatment, digital assets are property, which means every disposition is a taxable event, and a payment is a disposition. Buy coffee with a stablecoin, spend a token on a game item, purchase a domain with ETH — each is a realization event requiring cost-basis lookup and gain computation. Multiply that by a merchant accepting thousands of transactions per day.
If Congress sets a threshold that allows small gains to escape reporting, the payment and settlement sector gets its functionality back. If it does not, the tax friction is not a cost. It is a prohibition. No rational user computes basis on a four-dollar purchase. They stop using the asset for payments, and the utility thesis for the category decays by attrition rather than by headline. The provision nobody writes about is the one that quietly deletes a use case.
Question two: does the wash-sale rule extend to digital assets?
Equities have been subject to wash-sale rules for decades: sell at a loss, rebuy within thirty days, and the loss is disallowed. Digital assets currently sit outside that rule. Tax-loss harvesting has become one of the most reliable mechanical strategies in the market, and it is a large part of why December price action looks the way it does — sellers realizing losses, buyers re-entering immediately with no penalty.
Extend the rule and two things happen. The annual harvesting flow is compressed into different windows, reshaping year-end liquidity. And market makers who lean on harvesting as a component of inventory return lose that component. Their spreads widen to compensate. Depth thins. The effect is small per trade and large in aggregate, and it lands on the most liquid pairs first, which is to say it lands where hedging capacity matters most.
Question three: when is staking income recognized, and at what character?
Current guidance recognizes staking rewards as ordinary income at receipt, measured at fair market value when received. The recipient then holds an asset with a basis equal to that value, and any later sale produces capital gain or loss. If the asset falls forty percent before the recipient can sell, they owe income tax on a value that no longer exists.
This double exposure is the structural defect. It converts a capital decision into a cash-flow liability. Validators in jurisdictions that tax this way must liquidate a portion of rewards to cover the tax, which creates permanent sell pressure on every proof-of-stake network with resident operators. When the yield is too high, the exit is rigged. The headline staking rate means nothing until you compute the after-tax, after-liquidation-coverage return. For a high-bracket taxpayer in a drawdown, that number is frequently negative.
During the 2020 cycle I modeled liquidation cascades across the major lending markets and argued that DeFi had reproduced traditional finance's fragility with higher fees. The criticism was unpopular and, shortly afterward, correct. The mechanical point carries over directly here: a rule that forces periodic forced selling into a falling market is not a tax policy. It is a liquidation engine with a filing deadline.
Change the recognition point from receipt to disposal and the network's security budget changes, because the marginal validator's cost structure changes. That is a tax provision altering consensus economics. Legislators will not frame it that way. It is true anyway.
Question four: what basis method is mandated, and are cross-chain moves dispositions?
First-in-first-out versus specific identification is a technicality to a legislator and a six-figure difference to a trader. Beyond that lie the genuinely unresolved questions — the ones the technology created and the code has not answered.
Does moving an asset from a base layer to a rollup constitute a disposition? Does bridging through a third-party validator set? Does wrapping? Does depositing into a liquidity pool and withdrawing a different token composition, which is economically a sale of one basket for another even though no counterparty was ever involved?
I have done this reconstruction by hand. When I traced the wallet flows behind a mint that drained twelve ETH into offshore addresses within hours of launch, I had to assemble a transaction graph across five addresses, two chains, and one aggregator to establish a defensible timeline. I trace the wallet, not the whisper. That exercise took days and covered a few months of activity by a single actor. Scale it to every taxpayer with a self-custody wallet, a rollup account, and a bridge history.
This is where the rollup conversation becomes a tax conversation. The industry has spent years arguing about data availability and settlement layers without noticing that the same architecture generates the hardest basis-attribution problem in the code. Every message passed between layers is a potential realization event if the statute is written carelessly, and a chain migration that costs the user nothing in gas can cost them a capital gain in April.
If the statute mandates lot-level tracking without defining how a bridge or a layer-two migration maps to a taxable event, it is not creating clarity. It is creating an audit factory where the burden of proof sits on the user, and the resulting liability is unknowable until an examiner decides it.
Question five: who is a broker?
This is the whole bill. Everything else is decoration.
Reporting obligations attach to brokers, and the definition of broker determines who must file with the revenue service and, more consequentially, who must collect identity data in order to file. The 2021 infrastructure legislation contained a late-drafted broker provision whose language was broad enough that commentators argued it could reach software developers and validators. Guidance and subsequent legislative action narrowed it, and a Congressional Review Act resolution struck the rule that would have extended broker status to certain decentralized front-ends. Separately, the digital-asset information-return regime is phasing in for custodial intermediaries, with full basis reporting arriving after the first wave.
Watch three words in whatever text eventually appears: broker, digital asset, effective date.
If "broker" reaches non-custodial interfaces, the practical consequence is not paperwork. It is architecture. A front-end that must identify users and report gross proceeds to a government must hold user data. A protocol that must hold user data must have an entity that holds it. That entity becomes a subpoena target, a sanctions chokepoint, and a jurisdictional anchor. And a governance token that can upgrade the interface can be compelled to upgrade it in ways its holders never voted for. That is not a compliance cost. It is a decentralization tax, and it is paid in the one currency these protocols cannot print — credible neutrality.
If "digital asset" is defined to align with securities law rather than tax law, the two regimes will contest the same tokens under different tests, and the industry pays two sets of advisors to lose the same argument twice.
If the effective date is retroactive, every holder who relied on existing guidance inherits a liability they could not have priced. Retroactivity is rare in tax law and never advertised. It is discovered in the text, usually by the first taxpayer who receives a notice.
The rate is not the parameter. The rule is.
Here is where the industry's attention is misallocated. Capital gains rate changes move capital between jurisdictions. Reporting obligations move capital out of asset classes entirely.
Consider the rules as hidden parameters on token economics. A de minimis exemption restores payment utility. A wash-sale extension compresses market-making returns. A receipt-based staking tax suppresses participation and therefore network security spend. A clear statement that non-custodial protocols are not brokers removes a compliance discount embedded in every DeFi valuation. A defined treatment for airdrops converts a contingent liability into a recognized event, which would reduce the "airdrop equals tax bomb" fear that has suppressed a growth mechanism for three years.
None of those are price predictions. They are mechanism changes, and mechanism changes are the only thing in this industry that compounds. A statute is infrastructure. It does not trade.
Who actually captures the value
Follow the mandatory spend. Whatever the text says, if it increases reporting complexity, basis computation, and identity verification, three sectors capture that budget before any trader does: tax software, custodial infrastructure, and blockchain analytics vendors. Their revenue is the tax code's direct derivative, and it is uncorrelated with price.
The lobbying counterparties are also misidentified. The industry assumes it is negotiating against itself. It is not. It is negotiating against a tax preparation industry, an accounting profession with decades of institutional relationships, and broker-dealer compliance departments that already own the reporting rails and would prefer definitions that keep new entrants off them. Those constituencies have more Senate relationships than any protocol. The fight over the broker definition will be won by the side that files more comment letters, not the side that posts more threads.
A profile picture is not a shield against fraud, and no statute is a shield against competition. The EU has had a unified framework since 2024. Singapore levies no general capital gains tax. The UAE taxes personal income at nothing. Hong Kong connects the region's capital pool under a territorial system. Every month Washington spends debating definitions is a month those jurisdictions spend onboarding capital that was waiting on an American answer. The "global standing" line in the coverage is not a talking point. It is a measurable flow, and it runs one direction.
State tax never unifies. Jurisdictions with their own regimes will keep their own treatment, layered on the federal code, layered on local rules. Multi-layer compliance cost is permanent regardless of what passes.
What a defensible bill would contain
If the text, when it appears, does not touch the following, the markup was theater: a de minimis threshold for personal transactions; a defined basis method permitting specific identification; recognition of staking rewards at disposal rather than receipt; explicit exclusion of non-custodial software, validators, and node operators from the broker definition; a clear mapping of bridge and rollup activity to non-realization events; and a prospective effective date with no retroactivity.
Six items. Each is a sentence. Together they determine whether the United States has a digital asset industry in 2030 or merely a large population of former participants who file amended returns.
The bullish case is right, and it is being made for the wrong reason
The single largest unlock available to this market is not a rate cut. It is accounting certainty. Institutional allocation depends on categorization, not conviction. A fund must be able to state in writing that a position is a capital asset, that its basis is determinable, that its reporting is auditable, and that its compliance officer carries no personal exposure. Today none of those statements can be made cleanly, which is why the marginal institutional dollar has concentrated into a handful of assets with uncontested treatment. Any statute converting "unresolved" into "defined" is worth more to valuation than any provision inside it.
So the bulls are directionally defensible. Note what that position depends on, though: a statute becoming law. Not a markup. Not a hearing. Not a statement of intent. The bullish case is a bet on completion of a chain whose first link is a calendar entry, and this industry's habit is to price the entire chain at the first link. That is how the 2020 yield loops were priced. That is how a $60 billion algorithmic stablecoin was priced before it broke its peg — and that mechanism was visible in the code to anyone willing to read the seigniorage loop, which is exactly what my post-mortem documented afterward, alongside the regulators who arrived two years late with subpoenas instead of standards. Here the code does not exist yet. There is nothing to read.
There is one more possibility the coverage has not raised, and I consider it the most likely of all. This may not be a crypto bill. A committee with jurisdiction over the entire federal tax code does not schedule markups for narrow topics; it schedules packages. A markup containing crypto provisions is more plausibly a revenue vehicle with broader content, in which crypto is one clause among dozens. If that is the case, the crypto clauses are bargaining chips, and bargaining chips are the first thing cut in conference. Hype is the only asset in a vacuum mint.
The date that matters is not September 16. It is the day the text is published, which will arrive days or hours before the markup and without the framing that surrounded the announcement. Read it for three things: the definition of broker, the effective date, and whether the crypto provisions stand alone or sit inside a package. Everything else published this week was a forecast wearing a journalist's byline. Wait for the document.