The Four Sentences
Four sentences. That is the entire verifiable public record of what may be the most consequential crypto tax vote in Europe this quarter. The French National Assembly passed an amendment to the budget bill touching crypto taxation. The amendment targets "certain stablecoin swaps." A separate exit-tax proposal remains under deliberation. A separate self-hosted wallet declaration requirement remains under deliberation. No rate. No defined tax base. No effective date. No transitional regime. No impact assessment published.
Read those four sentences again, because everything the market thinks it knows about this event is smuggled in between them. The phrase "certain stablecoin swaps" is not a definition — in a tax code written in the most granular prose in Europe, it is a placeholder. And yet within hours the anglophone timeline had already filed the story under "France bans stablecoins," a headline that is wrong in every clause: France did not ban anything, did not name a rate, and did not finish legislating.
I have been doing this long enough to recognize what a placeholder inside a budget amendment actually is. It is a signal that the political decision has been made and the technical decision has not. In tax law, the technical decision is the entire event. The politics tells you the direction of travel; the drafting tells you whether anyone will actually be able to comply. Right now France has published the direction and withheld the drafting, and the gap between those two things is where every mispricing in this story lives.
What the Code Actually Says
To understand why this amendment matters more than its four-sentence footprint suggests, you have to know the architecture it is cutting into. France taxes digital assets — actifs numériques — under Article 150 VH bis of the Code général des impôts, at a flat 30% prélèvement forfaitaire unique: 12.8% income tax plus 17.2% in social contributions. That number is not the interesting part. The interesting part is the trigger.
Under the current regime, a taxable disposal occurs when crypto is converted into fiat currency, or used to pay for goods and services. Crypto-to-crypto swaps are not exempt — they are deferred. The cost basis carries forward. This was a deliberate administrative simplification, and it was a good one: it gave the tax authority a single, clean, observable gate. Money leaves the crypto perimeter through a bank, and the bank has a record. Everything inside the perimeter stays untaxed until it exits.
That single architectural choice created the arbitrage that this amendment exists to close. If you never touch a euro, you never realize. A French resident could rotate ETH into USDC in a stress event, park there for eight months, rotate back, and never generate a taxable event — because a stablecoin is, for French tax purposes, just another actif numérique. The fiat exit gate was elegant. It was also, in retrospect, the protocol all along — the flaw wasn't in enforcement, it was in the ontology. Once stablecoins became the actual medium of settlement inside crypto rather than a parking spot, "only fiat exits are taxable" stopped describing reality.
So place this against the European backdrop. MiCA, Regulation 2023/1114, splits stablecoins into e-money tokens and asset-referenced tokens and defines both as payment instruments. DAC8, the EU directive on tax administrative cooperation, applies from 1 January 2026 with first reporting in 2027. The OECD's Crypto-Asset Reporting Framework begins first exchanges in 2027. France legislating in the 2026 budget is not legislating in a vacuum — it is writing domestic implementing law ahead of a data pipeline that switches on eighteen months later. Paris's stated ambition, carried through the AMF's PSAN licensing regime and its MiCA transition posture, has been to displace London as the EU's crypto venue. That is the context. A jurisdiction that bet its financial-center strategy on being the most crypto-legible place in Europe is now drafting a rule that makes the most-used crypto instrument a taxable event.
The Reclassification
Here is the mechanical inference that the four sentences do not state and that almost nobody has drawn out.
Converting a stablecoin into euros is already taxable under current law. A stablecoin is an actif numérique; selling USDC for EUR realizes a gain or loss against cost basis. There is no gap there. Therefore an amendment targeting "certain stablecoin swaps" cannot be about the fiat exit — that path is covered. It must be about the path that is currently deferred: crypto into stablecoin. And if crypto-into-stablecoin becomes a taxable disposal, the legislature has made a qualitative decision that has nothing to do with rates. It has decided that a stablecoin is a euro wearing a ticker.
That is not a parameter adjustment. It is a paradigm change dressed as a technical amendment, and it is the only reading of the four sentences that is internally consistent. Which means the headline number everyone will argue about — 30%, or some surcharge, or some new band — is the least important variable in the entire file. The base is the event. Once you reclassify the instrument, the rate is just arithmetic layered on top of a decision that already moved the boundary of what counts as money.
This is where the phrase "certain stablecoin swaps" stops being sloppy drafting and starts being dangerous drafting. Three completely different things fit inside it, and they have wildly different consequences. First: crypto-to-stablecoin, the deferred path I just described, which is the only reading that requires new legislation. Second: stablecoin-to-stablecoin, USDT into USDC, EURC into USDC — same economic substance, different ticker, and if that is taxable you have created a taxable event with literally zero realized economics. Third: stablecoin-to-fiat, which is already covered and needs no amendment at all. The first is a genuine policy choice. The second is an accounting catastrophe. The third is a red herring. All three are legal under the phrase as written, and the National Assembly has not told us which one it meant.
The Asymmetry Nobody Reported
The four sentences bundle three proposals as if they were a package. They are not a package. They are three items at three completely different levels of legislative maturity, and treating them as a single tightening event is the central narrative trap in this story.
Take enforceability. A stablecoin swap executed on a licensed centralised exchange or through a MiCA-authorised crypto-asset service provider leaves a centralised audit trail. The DGFiP can request it. The infrastructure already exists, because KYC already exists, because the PSAN regime and then MiCA required it. That is a database query with a stamp on it. Now take the self-hosted wallet declaration. There is no intermediary. There is no counterparty. There is the taxpayer's honesty, a penalty schedule, and whatever chain-clustering heuristics the administration can afford to build. One of these is a filing requirement. The other is an honour system with a fine attached.
Reporting them in adjacent sentences creates a false equivalence, and the equivalence is doing real work — it makes the package feel inevitable, like a single wall advancing. It isn't. The stablecoin clause is administrable today. The self-custody clause is not administrable at all, and the distance between those two facts is roughly the distance between a law and a wish.
Then there is the technical problem that will actually determine whether this works: stablecoins are not interchangeable with each other, and the tax code is about to discover that. Rotating USDT into USDC ahead of a depeg scare is standard risk management. Rotating EURC into USDC to access a deeper pool is standard treasury operation. If either is a taxable disposal, the taxpayer is now obligated to maintain multi-currency, multi-chain, multi-wallet cost-basis ledgers with FIFO ordering across a dozen networks and an unbounded set of DeFi positions. The compliance cost stops being a percentage and starts being a headcount. When the cost of complying exceeds the tax owed, you haven't collected revenue — you've manufactured non-compliance and called it enforcement.
Where the Friction Bites
I spent three weeks in 2020 modelling Aave's liquidation cascades under stress, and the thing that modelling taught me had nothing to do with collateral ratios. Under a 40% drawdown scenario the binding constraint was never the health factor. It was the friction of moving collateral. The entire risk-management toolkit in a deleveraging event is: rotate the volatile asset into a stablecoin, repay the borrow, wait out the cascade, re-enter. That rotation is the circuit breaker. Tax every rotation and you have taxed the circuit breaker — you have priced the one action that keeps a leveraged position from becoming a liquidation.
The economic effect here is not a revenue effect. It is a lock-in effect, and lock-in effects are systematically under-modelled because they show up as an absence rather than a number. You do not see the rebalancing that didn't happen. You see a market that got quieter, thinner, and more fragile, and you attribute it to sentiment. The stablecoin velocity that underpins every DeFi operation — collateral switching, liquidation defence, cross-protocol arbitrage, treasury rotation — all of it runs through the instrument this amendment is aiming at. Raise the cost of each hop and you don't collect a slice of the flow. You reduce the flow.
Which is the Laffer problem in its purest form. Raise the rate on a mobile, voluntary activity and the activity relocates or stops; the base shrinks faster than the rate rises; revenue goes sideways or down while the compliance apparatus gets bigger. And there is a nastier variant lurking in the drafting. If "swap" is defined against gross converted value rather than net realized gain — which is what happens when legislators write placeholders and administrators fill them in under time pressure — the effective burden on a routine rotation can be a multiple of the nominal rate. A taxpayer who moves 100,000 USDC in and out of a position six times at par has realized zero. Under a gross-value reading, they owe tax on 600,000 of "disposals." That is not a tax rate. That is a transaction tax on existing, and it will be discovered by the first person who files.
And notice the contradiction France is walking into. MiCA defines stablecoins as payment instruments. The French tax code, if this amendment lands as I read it, defines them as taxable exit points. Those are opposite regulatory philosophies aimed at the same object, and France is the member state that fought hardest for the MiCA stablecoin title. Liquidity is just social consensus in code, and euro stablecoin liquidity is the thinnest, most consensus-dependent liquidity in the market — EURC and EURCV both run on a fraction of dollar-stablecoin depth. Apply swap friction to a thin book and the market makers don't pay the tax. They leave the tax jurisdiction. SG Forge, a bank subsidiary, issues EURCV from Paris. France is simultaneously incubating the euro stablecoin and taxing the operation that makes it useful.
The Ledger Problem
I want to flag one exposure that the market has almost entirely ignored, because it is narrative rather than financial and narrative shocks don't show up in a spreadsheet.
When I wrote my thesis on the Bored Ape Yacht Club in 2021, the conclusion that mattered wasn't about JPEGs. It was that the product was the narrative of exclusivity, not the asset. Apply that lens to Ledger. Ledger is France's most important crypto company, and its entire product is a sentence: not your keys, not your coins. That sentence is the brand. It is the reason a hardware wallet exists as a category. Now imagine the home jurisdiction of that company passing a law that requires residents to declare the self-hosted wallets they control. The company is not fined. Its revenue does not collapse. But its core marketing claim acquires a domestic asterisk, and asterisks are how narratives die — slowly, then all at once.
Shadows in the shard, light in the ape. The value is rarely where the volume is. Everyone is watching the rate. The rate is noise. The structural exposure is a French company whose product is sovereignty, operating under a French law that treats sovereignty as a reporting obligation.
And there is a second-order point buried in the self-custody proposal that I think is the actual intent. If you cannot effectively assess or collect tax on self-hosted assets — and you cannot, not without the intermediary that doesn't exist — then what does a declaration obligation produce? It produces data. A holder map. The classic progression of tax administration is: require the declaration first, build the baseline, tighten the assessment later. The revenue is a downstream product; the information asset is the immediate one. Read the self-custody clause as a data acquisition statute and its persistence makes far more sense than reading it as a revenue statute.
What the Market Is Misreading
Here is where I part company with the consensus framing, which is that France is cracking down.
France is not cracking down. France is surrendering — in a jurisdiction competition, quietly, while describing it as enforcement. You do not crack down on an asset class by raising the friction on holding it; you crack down by restricting access, and France has done the opposite for four years, building the most permissive licensing regime in Europe. What this amendment does is raise the carrying cost of a bearer asset in a world where the bearer can become a resident of Lisbon, Zug, or Dubai with a change of address and a tax return. In an asset class where the capital is maximally mobile, raising friction is not a levy on wealth. It is an export of the tax base and, eventually, of the people who hold it. Arbitraging culture before the code catches up cuts both ways: the French crypto community will process this culturally long before the code is final, and a meaningful slice of the largest portfolios will simply change jurisdiction while the drafting is still in committee.
The second misread is the legislative one, and it is the most expensive. "National Assembly passes" is being priced as "law." It is not law. The chain from here runs through the Sénat, whose finance commission has a long habit of rewriting crypto provisions on its own terms; then a joint committee to reconcile the two chambers; then the Conseil constitutionnel, which has a documented history of striking budget provisions on procedural grounds or on equality-of-taxation principles; and over all of it hangs the possibility of Article 49.3, the constitutional tool that lets the government force a budget through without a vote, at which point the fate of any given clause is decided by political arithmetic rather than technical merit. Current status is mid-legislation, not post-legislation.
And the three proposals will not travel together. My read, and I'll mark it as a read rather than a fact: the stablecoin clause survives in some form, because it is the only one with a data source behind it. The exit tax gets weakened or dies, because it carries the heaviest legal freight — France's existing exit tax bites on holdings above roughly 800,000 euros of equity, and extending it to crypto requires settling valuation timing, constitutionality against freedom of movement and property rights, and interaction with double-taxation treaties. The self-custody declaration gets delayed, diluted, or deleted, because its enforceability is an honour system and its constitutionality is genuinely contested.
The third misread is the direction of the behavioural response. The stated goal is visibility. The predictable outcome is opacity. If stablecoin conversion becomes a taxable event inside France, the rational French user does not stop converting — they convert on an offshore venue or through a peer-to-peer counterparty, where no CASP files a report. You have not captured the flow. You have pushed it outside the perimeter you built. Decoding the narrative before the fork happens means noticing that the enforcement action and the policy goal are pointed in opposite directions, and that the gap between them is where the actual market structure will reorganise.
What to Watch
Ignore the vote. The vote already happened and it is the least informative datapoint in the file. Watch three things instead, in order of information value.
First, the Sénat finance commission's text, because that is where a placeholder either becomes a definition or gets removed. Second, the definition itself: does "swap" mean disposal, or does it mean payment? Those two readings produce different tax bases, different compliance burdens, and different constitutional exposure. Third, the effective date — prospective from promulgation, or retroactive to 1 January 2026. Retroactivity would be the single most consequential clause in the entire bill, and it is currently invisible.
And then sit with the question the drafters have not answered. If a stablecoin is a euro for tax purposes, what is it for MiCA purposes? Two regulators, one instrument, two ontologies. When a French CASP has to file under one framework that calls the token a payment instrument and another that calls the same token a taxable exit, which description wins — and who absorbs the cost of the contradiction?