The vote was 38 to 37. Twenty-five abstained. A single legislator, Nicolas Sansu, carried Amendment No. 3898 through the French National Assembly by one ballot — and in doing so, drew a tax line not between profit and loss, but between two tokens that never left a wallet.
That is the entire event. No protocol launched. No chain halted. No balance changed. But a legislative body, voting on a 2027 budget rider, decided that converting Bitcoin into USDC should be treated as a taxable disposal event — a realization of capital gains at the moment of swap, even though the user receives no euros, no dollars, no payment instrument of any kind. Only a different ticker on the same screen.
Audit gap confirmed.
What the Existing Framework Actually Says
To understand what changed, you have to understand what already existed. Under current French treatment, crypto-to-crypto swaps sit in a deferral regime. If you move from BTC to ETH, no taxable event is recognized. The gain is unrealized. The obligation is deferred until you convert into fiat currency — at which point the tax authority takes its position and the accounting settles.
That deferral is not a loophole. It is an accounting reality. A swap between two volatile assets does not produce a fixed, measurable gain because the receiving asset may appreciate or depreciate before any conversion. The gain is a moving target. France, to its credit, recognized this and deferred.
The stablecoin carve-out breaks that logic. Stablecoins, in the eyes of the amendment's sponsor, are not really crypto assets. They are quasi-cash. They track the euro or the dollar, they hold value, they sit in wallets as dry powder awaiting deployment. Sansu's argument, as reported, is straightforward: stablecoins should not receive better tax treatment than the currency they imitate.
It is a clean argument. It is also mathematically incomplete.
The Category Error at the Center
Here is where the forensic reading matters. A stablecoin is designed to track fiat. It is not guaranteed to. The distinction is not academic — it is the entire historical record of the asset class. USDC traded at 0.87 during the March 2023 banking crisis. USDT has traded through discount windows repeatedly. Algorithmic designs have gone to zero. Even the largest, most liquid stablecoins have experienced depeg events measured in hours, not hypotheticals.
If you legislate that a stablecoin equals cash, you are legislating away the tail risk. And when that tail risk materializes — as it has, repeatedly, across multiple issuers and multiple cycles — the taxpayer bears the consequence of a definition the state wrote for its own convenience.
Consider the position of the French resident who swaps 1 BTC acquired at 30,000 euros for USDC at a market value of 60,000 euros. Under the amendment, that is a 30,000-euro taxable gain. The asset received is USDC — nominally 60,000 euros of purchasing power. But the user has received no euros. To satisfy the tax obligation, they must either sell part of the USDC for fiat, or fund the tax from external income. The state has taxed a gain the user cannot spend without first converting the very asset the state just declared equivalent to cash. That is not a tax on profit. It is a tax on the act of repositioning, financed from outside the transaction.
Yield trap detected — except this time the trap is fiscal, not incentive-driven.
The Vote Structure Tells You More Than the Vote Outcome
38 to 37 is the headline. The more informative number is 25.
In a chamber where budget amendments are routine, a near-tie with a quarter of participants abstaining is not a mandate. It is an expression of avoidance. Abstention on a tax question is unusual — legislators typically prefer to be recorded on either side, because tax votes are politically legible. When 25 members decline to take a position, the signal is that the issue is contested within their own coalitions, and no one wishes to own the outcome.
That structure has predictive value. French budget amendments, in normal practice, do not survive the full legislative path intact. They pass the National Assembly, meet the Senate, and then enter the Commission Mixte Paritaire — the joint committee where divergent texts are reconciled. Historically, that stage is where narrow-margin riders are diluted, reframed, or removed entirely. A single-vote amendment with 25 abstentions is precisely the profile of a provision that gets reconstructed downstream.
Mathematical collapse verified — not of the market, but of the legislative certainty.
I have watched this pattern before. In 2017, I audited fifteen ERC-20 contracts during the ICO peak and found reentrancy vulnerabilities in three prominent projects. The code was verifiable. The conclusion was inevitable. But the community response was not about the code — it was about whether the finding should have been published at all, given the timing. The lesson I took was that verifiable defects do not become operative until someone acts on them, and the distance between detection and enforcement is where most of the actual risk lives.
The same principle applies here. A one-vote margin in a budget rider is a detected defect in the tax architecture. It is not yet an enforced liability. Any French holder who restructures their positions today, ahead of a final text that may never materialize in this form, is pricing in certainty that the vote count does not support.
What the Current Law Does Not Address
The amendment, as reported, is silent on the operational questions that determine whether it is enforceable at all.
There is no stated methodology for valuation. Is the gain computed on the fiat-denominated value of the stablecoin received at the moment of swap? On the cost basis of the asset surrendered? On a weighted-average basis across prior acquisitions? France has an existing cost-basis framework for crypto disposals, but whether it maps cleanly onto intra-wallet swaps is not established.
There is no stated position on retroactivity. Does the amendment apply to swaps executed after enactment only, or does it reach backward into prior tax years? This is not a technicality. For any holder who has been rotating between volatile assets and stablecoins across multiple cycles, retroactive application would require a full reconstruction of historical positions, with valuations tied to exchange rate data that may no longer be retrievable from a defunct venue.
And there is no stated position on the enforcement asymmetry. Centralized exchanges operating in France under registration obligations can be compelled to report. Self-custodied wallets cannot. A tax regime that is enforceable only against users of registered intermediaries, while remaining structurally unenforceable against self-custody, produces a differential burden between compliant and non-compliant users. The compliant user pays; the non-compliant user does not; the incentive gradient points away from compliance.
Based on my audit experience, this is the recurring failure mode in crypto tax design. Legislatures write rules that assume data availability, then discover that the data exists only for the subset of activity that already reports voluntarily.
The Bull Case Nobody Is Making
There is a reading of this amendment that its opponents are missing, and it deserves a hearing.
If France eventually concludes that stablecoin swaps are taxable disposals, it will have implicitly accepted that stablecoins are a distinct asset class requiring distinct treatment. That is not a hostile position — it is a classification. And classification is the precondition for every serious regulatory framework, including MiCA, which already separates e-money tokens and asset-referenced tokens from other crypto assets.
The bull argument is that clear, if burdensome, rules beat ambiguous ones. A French holder who knows that BTC-to-USDC is a taxable event can plan accordingly. A holder operating under an ambiguous deferral regime, uncertain whether a particular swap structure will be recharacterized in an audit, cannot. Certainty has value even when the certainty is expensive.
The counter is that this certainty is not yet certain. The amendment has passed one chamber by one vote with a quarter abstention. Until it survives reconciliation, the only genuine certainty is that the rule does not exist.
Ledger does not lie: the record shows a draft, not a statute.
The Signal Beneath the Noise
The substance of Amendment No. 3898 is minor in market terms. France is one jurisdiction. The affected population is French tax residents with stablecoin activity. There is no fiscal estimate attached to the proposal, no user-scale data, no projected revenue figure — all of which are material gaps in any serious tax legislation.
What matters is the direction of the reasoning. A legislature has now formally debated whether the boundary between crypto trading and taxable realization should be drawn at the stablecoin, rather than at fiat conversion. That boundary question is not uniquely French. It is the same question being asked in every jurisdiction that has accepted crypto assets as property. The answer France eventually gives — through reconciliation, through Senate revision, through the final text — will be read as a precedent.
And the tax compliance industry will read it first. Whatever the outcome, the distinction between crypto-to-crypto and crypto-to-stablecoin has been pulled into the light. Every French-facing exchange now has a reason to build reporting infrastructure that can separate those two flows. Every tax software vendor has a reason to model both. That is a real, if modest, reallocation of engineering attention, and it will happen whether or not the amendment survives.
The question worth asking is not whether France will tax stablecoin swaps. It is whether any legislature has yet demonstrated the ability to define a stablecoin precisely enough to tax it — and whether the definition will hold when the next depeg event tests it.
One vote is not an answer. It is a question with a margin.