On a Tuesday in the French budget cycle, a number appeared that most crypto media filed as a footnote: 800,000 euros. That is the threshold at which a French tax resident holding crypto assets would, under a newly floated proposal, trigger an exit tax upon relocating their fiscal domicile abroad. The figure is not arbitrary. It maps exactly onto Article 167 bis of the Code général des impôts, the existing French exit tax regime for securities. The ledger remembers what the hype forgets. This is not a new tax. It is a copy of an old one, with the asset-class variable swapped from securities to crypto-assets. That distinction matters more than the headline number, because the mechanics of the old regime — deemed realization, deferral, guarantees — are the actual attack surface, not the rate.
France has spent the last four years positioning itself as the European Union's regulatory anchor for crypto. It was among the loudest advocates for MiCA, the Markets in Crypto-Assets regulation that now governs the bloc's licensing regime. It pushed for DAC8, the eighth iteration of the Directive on Administrative Cooperation, which obliges member states to automatically exchange tax information tied to crypto-asset holdings and transactions. And it sits inside the OECD's Crypto-Asset Reporting Framework, or CARF, the cross-border reporting standard built to make crypto holdings legible to tax authorities in the same way bank accounts already are.
Read those three instruments together and the exit tax proposal stops looking like an isolated budget grab. It becomes the enforcement arm of a system assembled piece by piece. MiCA sets the licensing perimeter. DAC8 builds the reporting pipes. CARF standardizes the data format across jurisdictions. The exit tax is what happens when a holder tries to walk out the door before the pipes are fully connected.
The French logic is straightforward, and from a revenue standpoint, rational. A tax resident holding more than 800,000 euros who relocates to a zero-tax jurisdiction is, in the state's view, exporting an unrealized tax liability. The exit tax treats relocation as a taxable event. The gain not yet crystallized is deemed realized at the moment of departure.
This is not novel fiscal engineering. France has applied the mechanism to securities for years. What is novel is the extension to self-custodied, pseudo-anonymous, cross-border-transferable assets. That is where the mechanics begin to strain against reality.
The Mechanics That Decide Everything
The core of any exit tax is the definition of the taxable event, and the proposal as reported leaves three mechanisms unspecified. First, is the gain treated as deemed realization, with tax immediately due on departure? Second, is collection deferred until the asset is actually sold? Third, is the holder required to post a guarantee or file a declaration to secure the deferred liability? These are not semantic distinctions. They determine whether the policy is a paper tiger or a genuine constraint.
The existing Article 167 bis regime leans on a combination. Collection is deferred, with the possibility of an exemption if the holder moves within the EU/EEA and holds the asset for a defined period, or if the destination jurisdiction maintains adequate information-exchange arrangements. Whether the crypto version replicates that deferral is the single most important unknown in the proposal. My confidence that it does is moderate — the 800,000 euro threshold match is too precise to be coincidence.
The historical parallel is instructive. The United States operates an expatriation tax under the HEART Act, which taxes "covered expatriates" on the deemed sale of their worldwide assets upon renouncing citizenship or long-term residency. Germany applies an exit tax on substantial shareholdings under Section 6 of the Foreign Tax Act. Both regimes share a common architecture: a net-worth or ownership threshold, a deemed-realization trigger, and a set of deferral or exemption mechanisms that soften the blow for compliant movers. France is not inventing a category. It is extending an established category to a new asset class.
That extension is where the engineering fails, and this is where my audit background becomes relevant. I have spent the better part of a decade reading smart contracts and, more recently, the compliance architecture around them. The pattern I keep encountering is that policy designers assume the on-chain world is as legible as a brokerage statement. It is not. A French tax resident's crypto holdings are not registered anywhere by default. They can sit in a hardware wallet, behind a seed phrase, inside a multi-signature contract, or on a self-hosted node. The state cannot query a database to establish the 800,000 euro threshold. It can only rely on self-declaration and, eventually, on third-party reporting.
This is the same logic gap that runs through most crypto regulation. Logic gaps leave holes in the smart contract — and in the tax code. The exit tax assumes a verifiable asset base. The asset base is only verifiable if the holder cooperates. Cooperation is precisely what a holder contemplating relocation is least inclined to provide.
The Enforceability Gap
Consider the reporting layer meant to close that gap. DAC8 requires crypto-asset service providers — exchanges, custodians, brokers — to report the holdings and transactions of EU residents to their tax authorities. CARF extends that reporting across OECD borders. On paper, these instruments make the exit tax enforceable for anyone who touches a regulated intermediary. If a French resident holds Bitcoin on a licensed exchange, the exchange reports it. If they sell on a compliant platform after moving to Dubai, the platform reports that too, and the data flows back to France.
The problem is that the most mobile, most valuable holders are the least likely to sit inside that perimeter. Self-custody is not an edge case; it is the default posture for high-net-worth crypto holders, and for good reason. Every major custodial failure of the last cycle — exchange insolvencies, frozen withdrawal queues, seized accounts — taught the same lesson. Trust is a variable, not a constant. A holder with eight figures of exposure has already internalized that. They hold their own keys. An asset under a seed phrase generates no third-party report.
So the exit tax's enforcement rests on a behavioral assumption the target population is structurally disinclined to satisfy. This is not unique to France. It is the central tension of every attempt to tax self-custodied assets. The state wants to treat crypto like a bank account. The asset is designed not to be one.
In 2017, I spent forty hours manually auditing the Solidity of an ICO that promised decentralized cloud storage. The whitepaper described a functioning network. The contract described an integer overflow in the minting function. I found it with a custom Python script, reported it by email, received no reply, and published the breakdown. The lesson was not that the team was malicious. It was that the gap between the stated system and the implemented system is where all the risk lives. The same gap exists here. The stated system is "we will tax unrealized gains on departure." The implemented system — pending — will be some combination of declaration, guarantee, and information exchange. The distance between those two is where every enforcement failure will occur.
The Pull-Forward Problem and Jurisdiction Competition
There is a second-order effect the authors may not have fully priced: the pull-forward dynamic. Any credible signal that an exit tax is coming creates an incentive to leave before it takes effect. If the proposal moves toward legislation, French residents above the threshold face a simple calculation: relocate now, while the regime is a draft, or relocate later, under a defined liability. Rational actors with the means to move will front-run the effective date. The policy may therefore accelerate the very tax-base erosion it was designed to prevent — at least in the near term, before any transition period or guarantee requirement binds.
The second channel is jurisdiction competition. Tax regimes compete for mobile capital, and crypto wealth is among the most mobile capital ever created. The proposal hands a relative advantage to jurisdictions with no exit tax and no capital gains tax — the UAE, Singapore, and a rotating cast of others. If the French regime lands in a strict form, the marginal high-net-worth holder does not argue; they move. The wealth-management, tax-advisory, and relocation-services industries absorb the demand. French exchanges and custodians absorb the loss. The winners and losers of this policy are not where the headline points.
The third channel is the compliance-tooling market. Every expansion of reporting obligations creates demand for software that satisfies them. On-chain analytics, portfolio tracking with tax-export functionality, jurisdiction-aware reporting dashboards — these become line items. The proposal, if it advances, is a subsidy to that sector. This is the same dynamic I observed when I audited the AI-agent trading platform in 2025 and found a reentrancy vulnerability in its cross-chain bridge. The platform's promise of autonomous yield generation was underwritten by a contract that could be drained. The compliance market that grows around a tax regime is underwritten by a similar assumption: that the regime is enforceable. If it is not, the tooling still sells, because the obligation to report exists whether or not the underlying data is complete.
The Coordination Effect
France is not acting alone. The exit tax is one instrument in a portfolio that includes DAC8 and CARF. Together they are designed to make "move and do not declare" a losing strategy. DAC8's automatic exchange means that if a holder uses a regulated intermediary in the destination country, the data returns home. CARF extends that logic across the OECD. The combined effect is to shrink the space in which relocation is tax-effective.
This is the part the market may be underpricing: not the exit tax in isolation, but the exit tax as the enforcement edge of a coordinated transparency regime. A single country's exit tax is easy to route around. A cross-border reporting framework with automatic exchange is harder. The proposal's real weight comes from its position in that stack, not from its own text.
There is a caution here that applies beyond tax policy. Regulators, like infrastructure builders, tend to over-engineer for demand that may never materialize. We have watched this in the rollup space, where dedicated data-availability layers are provisioned for throughput that the vast majority of chains will never generate. The same instinct shows up in tax enforcement: elaborate reporting machinery built for a data flow that self-custody quietly bypasses. Capability is not the same as capture.
What the proposal does not change is equally important. It does not alter the supply of any token. It does not change a single line of protocol code. It does not touch the mechanics of any DeFi contract. From a market-structure standpoint, its direct price impact is negligible. What it changes is the after-tax calculus of a specific, small, geographically concentrated population: French tax residents holding more than 800,000 euros in crypto assets who are considering relocation. For that cohort, the proposal is material. For the broader market, it is noise dressed as signal.
The Enforcement Asymmetry
Tax policy assumes symmetric information: the state knows what you hold, and you know what you owe. Crypto breaks that symmetry in one direction. The holder has complete information about their own position. The state has almost none. Every enforcement mechanism the proposal can deploy is an attempt to close that asymmetry from the state's side — through reporting obligations on intermediaries, through self-declaration requirements, through guarantees. None of them restore symmetry for assets that never touch a regulated intermediary.
The empirical record supports a cautious reading. Crypto tax enforcement has consistently lagged the legislation that authorizes it. Reporting regimes take years to phase in. Exchange data arrives incomplete and in inconsistent formats. Jurisdictions implement CARF on different timelines, so the "automatic" in automatic exchange is automatic only where both sides are ready. A holder can relocate during the gap.
I watched a version of this play out in 2020, when I reverse-engineered the Compound Protocol's interest rate model during the DeFi Summer run. The reported TVL and the actual on-chain collateral utilization rate diverged. The discrepancy was invisible to anyone reading the dashboard and obvious to anyone reading the contract. The lesson carried forward: the number people quote and the number that matters are rarely the same. The same will be true of the exit tax. The quoted rate will dominate the headlines. The effective rate — shaped by deferral, exemption, guarantee, and the destination jurisdiction — will be the number that actually binds.
For a French holder above the threshold, the practical question is not whether the tax exists. It is whether the state can establish the fact of the holding at the moment residency changes. Everything else is arithmetic on top of an unverified base.
The Real Risk Is the Definition, Not the Rate
Everyone is watching the rate and the threshold. The blind spot is the definition of tax residency and the disclosure regime that will accompany enforcement.
An exit tax is only as strong as the state's ability to identify who owes it. That requires two things: a clear definition of when residency ends, and a mechanism to discover holdings at the moment of departure. The first is a legal question with decades of precedent. The second is a surveillance question with almost none.
Here is the contrarian read. The exit tax may not be designed to collect much revenue at all. France can model the likely yield, and the modeling almost certainly shows a modest number — the affected population is small, the enforcement gap is wide, and the pull-forward effect erodes the base. The proposal's function may be deterrent-by-announcement: a signal that relocation is being watched, and that the window for frictionless departure is closing.
If that is the intent, the real risk is not the tax bill. It is the normalization of a premise: that self-custodied crypto is inherently suspect, and that holding it without disclosing it to the state is a compliance failure waiting to be defined. That premise, once legislated, is portable. It is the same premise that turned open-source developers into defendants in the Tornado Cash matter — the reach for an instrument built for custodial, jurisdiction-bound intermediaries and its application to an architecture that is none of those things. The instrument does not fit the object. The result is a policy broad in stated ambition and narrow in practical reach, but durable in precedent.
This is why I would watch the disclosure mechanics more closely than the headline rate. If the crypto version of Article 167 bis imports a guarantee requirement — a lien, a blocked account, a third-party pledge — then the state has found a way to reach into self-custody indirectly, by taxing the act of leaving rather than the asset itself. That is the mechanism that would actually bite. The rate is theater. The guarantee is the contract.
Takeaway
The proposal is a proposal. It is not law. Its final form — deemed realization versus deferral, the presence or absence of a transition period, the treatment of pre-existing holdings, the retroactivity question — is unknown. The bug was there before the launch, and in this case the launch has not happened.
Three signals will decide whether this is a footnote or a template. First, whether the text specifies a guarantee or declaration mechanism, because that determines enforceability against self-custody. Second, whether the deferral and exemption structure mirrors Article 167 bis, because that determines how hard the regime actually bites for compliant movers. Third, whether other jurisdictions copy the model, because that determines whether the exit tax becomes a category rather than a curiosity.
The pattern is consistent across every cycle I have audited: policy arrives first, enforcement second, and the market prices neither until the second shoe drops. The exit tax is the first shoe. The guarantee mechanism, if it comes, is the second. Watch the second.
Clarity precedes capital; chaos precedes collapse. Right now there is neither. There is a draft, a threshold, and a reporting framework being assembled around it. The market will not price any of it until the draft becomes text. By then, the actors who were going to move will already have moved. Data does not lie; people do — and the ones with the most to lose are already reading the fine print.