Korea Has a Tax Date and No Definition

Fifty-three thousand eight hundred and nine. That is the number of signatures a petition to abolish Korea's crypto income tax gathered in May. In mid-September, a second petition — this one asking for a two-year delay — crossed the same 50,000-signature threshold that forces National Assembly review. Two thresholds in four months. That is not sentiment. That is a queue. But the figure I keep returning to is not a signature count. It is a date: January 1, 2027. Korea intends to begin taxing gains on an asset class whose legal definition, by the admission of the regulator pushing the bill, has not yet been written. Lee Eok-won, chair of the Financial Services Commission, says the Digital Asset Basic Act is "close to completion" and will be pushed to the National Assembly as soon as possible. Three pillars are on the table: licensing requirements for exchanges, issuance rules for stablecoins, and a clearer legal definition of what a digital asset actually is. That is the entire visible structure. The capital requirements and reserve rules that would determine who is permitted to issue a won-denominated stablecoin remain undisclosed and, per the same reporting, still subject to change. Meanwhile the National Tax Service is running its own clock. A draft is expected by October 30, with public notice and initial review in November. The tax itself — 20% on annual gains above 2.5 million won, 22% including local surtax — has been rescheduled three times: from 2022 to 2023, to 2025, and now to 2027. I should flag my own evidence here, because the tightest dates in this story — the October 30 draft, the November review — trace back to an anonymous crypto account rather than an NTS filing. The timeline is plausible. It is not verified. In my fund's notes that gets a yellow tag, not a red one, but it does not get promoted to a fact. Ruling party leader Kim Min-seok, asked about the tax, said he is reviewing various opinions and called for a thorough final review. In political language that is not a position. It is a reservation. Here is the structural problem, and it is not political. Korea is preparing to levy an income tax on an asset whose property boundary is undefined. In legal terms that boundary decides whether a token is a security under the Capital Markets Act, a digital asset under the new act, or some third thing. In engineering terms it decides everything downstream: which token standard a compliant issuer uses, whether transfers require a whitelist, whether the contract enforces transfer restrictions at all. You cannot finalize a contract architecture against a definition that does not exist. And you cannot compute a capital gain against a cost-basis regime that has not been specified. I learned this the tedious way. In 2017 I spent three weeks modeling Golem's reward distribution against transaction-fee volatility and found the schedule held only if fees stayed inside a band nobody had specified. The Korean tax calendar has the same shape. It is not a claim about time. It is a claim about incentives. If the incentive to postpone persists, the date is decoration. That sequencing failure has a measurable economic signature. Realization-based taxation — taxing the year you sell rather than the year you accrue — does two things. It suppresses turnover, because every disposal becomes a taxable event and the rational holder defers. And it manufactures an industry: cost-basis tracking, loss harvesting, jurisdiction-aware reporting. Every tax regime creates the plumbing it needs in order to be paid. Korea has announced the pipe before drawing the plumbing. Whether this actually binds depends on variables nobody has published. Rate is known. Exemption is known. What is unknown: how cost basis is computed, whether staking and lending yield fall in scope, whether foreign holders are treated differently. Those four unknowns determine the entire incidence of the tax. Without them, anyone claiming to have quantified the market impact is selling you a model with the coefficients deleted. Math does not care about your conviction, and it does not care about a press conference. The stablecoin silence is the tell. Exchange licensing is a largely solved design problem here — the 2021 reporting regime already exists, and the upgrade to licensing adds mainly verifiability of client asset segregation, Travel Rule data interfaces, and transaction monitoring. Necessary, expensive, not mysterious. Stablecoin issuance is where the real negotiation sits, because eligibility determines who captures the won-denominated float. A bank consortium and a platform company want different capital rules. When a regulator announces the existence of a rule and withholds its content, that is usually not secrecy. It is an unfinished argument. And the definition outranks both, because it is the ground the other two pillars stand on. Narratives are liquid; truth is solid. The definition is the solid part. Everything else in this bill is a layer above it. The popular reading of this kind of headline is that clarity is bullish. I would separate direction from content, because they trade at different prices and settle on different dates. Directional clarity — "a framework is coming" — is already in the price. Korean exchanges have operated under a reporting regime for four years; institutions have been waiting for the legislative layer, not discovering it. That repricing happened quietly, over months, not in a headline. Content clarity — which entities may issue stablecoins, what reserves back them, where the security line falls — is not priced, will not be priced until the text is published, and may not survive the first reading in the Assembly. In the chaos, look for the invariant. Here it is: collection requires definition, and definition is the one thing that has not been delivered. The genuinely contrarian point is that the bill is not the sensitive variable. The tax is. Three delays have trained participants to expect a fourth, which means any "confirmed" date now gets discounted toward zero — including a genuine commitment. Quietly positioned while the world shouts: the people who will be right here are not forecasting the vote, they are modeling the incentive to postpone. The revenue at stake is modest. The political cost of collecting it is not. That asymmetry already produced three delays. It is still there. One more blind spot. The bill regulates intermediaries — exchanges and issuers — and says essentially nothing about DeFi, DAOs, or NFTs. That vacuum will be read online as provisional permission. It is not permission. It is undefined space, and undefined space is where compliance risk runs highest. Watch November, not January 2027. If the initial review produces text, the definition and the reserve rules will be the two sentences that move capital; everything before them was theater. If it produces another postponement, Korea has not built a framework — it has built a calendar. The question was never whether Korea legislates. It is whether a state can collect before it defines, and what it tells us about a market that allows it.