A 51% markup on Nasdaq just taught crypto’s oldest lesson to the stock market.
In the second week of July, a Korean office worker did something that looks, on paper, insane. She wired money to New York to buy shares of SK Hynix — a company headquartered forty minutes from her desk, listed on an exchange she can access with a tap — and paid 51% more than the Seoul price. She was not alone. That week, Koreans made the New York line of their own national champion the second-most-bought foreign stock in the country, half a billion dollars of it, while the Seoul line sagged.
Veterans of an older mania will recognise the shape. In 2017, Bitcoin in Seoul traded as much as 50% above Bitcoin everywhere else. The gap got a name — the kimchi premium — and became the world’s shorthand for Korean retail fever. Economists wrote papers about it. Regulators cited it. Then, as far as most people know, it died.
It didn’t. We measure it every thirty minutes, across eight pairs, and July’s ledger tells a duller, stranger story: for anything with depth — Bitcoin, Ethereum, Solana — the premium never left a band of a tenth of one percent. Only the illiquid meme coins strayed, by whole percentage points, which is exactly what you’d expect where depth is thin. Through the won’s biggest monthly move in seventeen years, through the first coordinated dollar-selling by Seoul, Tokyo and Washington in history, the famous premium on the majors was a rounding error.
Why is the premium 500 times larger on a Nasdaq-listed chipmaker than on Bitcoin, the asset that made it famous?
The answer is plumbing. A coin crosses between a Korean and an American exchange in minutes; arbitrage eats any gap before lunch. A Hynix ADR cannot be converted into a Seoul share and flown home — no pipe exists. Fever is constant; friction is not. The same Korean buyer who closes a 0.1% Bitcoin gap in real time will hold a 51% Hynix gap open for days, because holding it open is all she can do.
Which means she wasn’t crazy at all: she was simply the first to price in that nobody could arbitrage her.
Two prices arguing
Our ADR ledger caught those days in motion — five readings, four sessions: 23%, 51%, 38%, 42%, 22%. Next to each reading we logged which side moved. New York overshot (+27% in a session); Seoul chased (+9%); Seoul stumbled (−12%); New York gave it back (−14%). At no point was one correct price being discovered. There were two prices arguing, and the ledger names whose turn it was.
The ADR was the loud part. The quiet part is what we measure every session: how American chip moves transmit into Korean ones. Ten years of data puts the pass-through at a beta of 0.26; the last 250 sessions put it at 0.47 and strengthening — Seoul’s chip complex is more coupled to New York now than at any point in our sample, with the US signal calling Korea’s direction 63% of the time across 1,643 sessions. In the very week of the listing, Korean semis fell a full point further than that relationship implied. A measured relationship, not a forecast: when Seoul keeps outrunning the New York signal, the reading fails — and we log the failure too.
The old premium, meanwhile, has begun to stir. As the won surged on intervention, our kimchi ledger flipped from a month of shallow discount to positive — the first flip in three weeks, peaking at a quarter of a percent on Friday. Small. Visible only if you happen to be writing it down every half hour.
A premium was never a property of Korean investors, or of crypto. It is the going rate for friction between two markets — and that rate is quoted daily, whether or not anyone is keeping the ledger.
What we publish here are concluded readings: premium levels, transmission betas, deviation points, sample counts. What we sell per call is the live ledger — kimchi premium across eight pairs every thirty minutes, funding and regime state, and session-by-session transmission with its falsifier conditions. Exchange pairs, ticker mappings and estimation design stay private. Four sessions and one listing are a boundary of observation, not a law.
Data: our own half-hourly premium ledger and daily transmission ledger; public exchange prices. Findings may be cited with attribution to Asia Pulse (asia-pulse.com).