Korea publishes preliminary trade data for the first ten and first twenty days of every month — typically on the 11th and the 21st, with the full prior month on the 1st — making it one of the earliest official readings on global trade available anywhere. The single most useful habit for reading it: skip the headline year-on-year growth rate, divide by working days first, and only then look at the semiconductor line. Because Korea’s memory-chip makers sit at the center of the global electronics supply chain, that one line is a near-real-time proxy for worldwide tech demand, published well ahead of most other major economies’ official trade figures.
Markets treat it accordingly. The recurring shape of a strong release day runs like this: the semiconductor export line beats, Samsung Electronics and SK hynix rally at or shortly after the open, the index reverses early weakness on their weight, and money changes hands — foreign investors covering light positioning while individual investors take profits into the strength. That pattern — a chip-led export beat, an index reversal, and a hand-off between foreign and retail money — repeats often enough that the release schedule belongs on every Korea watcher’s calendar. This guide explains what is in each release, how to adjust the numbers, and how the market actually trades them.
What the Korea Customs Service publishes, and when
The Korea Customs Service releases preliminary trade statistics three times a month, usually in the Seoul morning. When the scheduled date falls on a weekend or holiday, the release slips to the next business day.
| Release date | Coverage | What it tells you |
|---|---|---|
| 11th | Days 1–10 of the current month | The first hard-data read on the month; the noisiest of the three |
| 21st | Days 1–20 of the current month | More stable; usually anchors expectations for the full-month figure |
| 1st | Full previous month | Confirms or revises the picture the interim prints sketched |
Each release contains total exports and imports in US dollars with year-on-year comparisons, the trade balance, and — critically — breakdowns by major product category and by destination. The product lines include semiconductors, petroleum products, passenger cars, ships, wireless communication devices and steel; the destination lines cover China, the United States, the European Union, Vietnam and Taiwan, among others. Two lines carry most of the market-moving information: semiconductor exports (a demand signal) and, on the import side, semiconductor manufacturing equipment (a capex signal — Korean chipmakers importing more tools is an early hint that a memory upcycle has legs).
Why the chip line is the number that moves markets
Semiconductors are Korea’s largest export category, historically around one-fifth of total exports, and the bulk of that is memory — DRAM and NAND — produced by Samsung Electronics and SK hynix, which together dominate the global memory market and rank among the largest weights in the KOSPI. Memory is closer to a commodity than most chip products: prices swing with the global cycle, and export value is simply price times volume. So when the 10-day semiconductor export line accelerates, it means some combination of memory prices firming and shipment volumes rising — a direct, dollar-denominated read on data-center, smartphone and PC demand worldwide.
The timing is a large part of its value. Korean customs data for the first ten days of a month is public before most companies in the supply chain have said anything about that month at all. Global investors use it as an advance read not just on Korean earnings but on the whole memory complex and, by extension, tech demand broadly. A practical reading order for the release: total exports year-on-year, then working days (see below), then the semiconductor line, then chip-equipment imports, then exports to China — the largest single destination for Korean chips and the line where geopolitical and supply-chain shifts show up first.
The working-day trap: headline YoY versus daily average
The most common mistake with this data is quoting the headline year-on-year change without checking working days — joeop ilsu in Korean. Customs counts weekdays fully and Saturdays as half a working day, and the count for the same calendar window can differ meaningfully from one year to the next because of where weekends and holidays fall. The distortion is worst around the two big lunar holidays: Seollal (Lunar New Year, in January or February) and Chuseok (the autumn harvest holiday, in September or October), both of which shift between months from year to year and can swing an entire month’s comparison.
The fix is arithmetic, not judgment: divide total exports by the number of working days and compare the daily average year-on-year. A hypothetical shows why this matters. Suppose the first ten days of a month contain 7.5 working days this year versus 9 last year. Total exports could fall 10% year-on-year while daily-average exports rose 8% — the headline says contraction, the adjusted number says acceleration, and the adjusted number is the one that describes demand. The customs release itself states the working-day counts and usually the daily-average growth rate; wire headlines often lead with the unadjusted figure. When the two diverge, trust the daily average, and expect the market to figure this out within hours even if the first headline reaction goes the other way.
How the market trades the print: a worked example
Samsung Electronics and SK hynix typically react at or shortly after the Seoul open on release days, and the index follows because of their weight. But the more informative context is positioning — what foreign investors were doing in the sessions before the print. Foreign flows are published daily for KOSPI and KOSDAQ, quoted in eok won (units of 100 million won).
Here is a real sequence, from completed sessions in August 2026, covering the trading days immediately before a mid-month 10-day release:
| Session (2026) | Foreign net flow, KOSPI+KOSDAQ (eok won) |
|---|---|
| Aug 3 | −28,220 |
| Aug 4 | −3,716 |
| Aug 5 | +14,464 |
| Aug 6 | −32,893 |
| Aug 7 | −8,651 |
| Aug 10 | −14,909 |
Reading it step by step: sum the six sessions and foreigners were net sellers of roughly 73,900 eok won — about 7.4 trillion won, or on the order of US$5.3 billion at the won’s early-August 2026 level of around 1,400 per dollar. That is heavy selling into the print, which tells you foreign positioning in Korean equities was light and expectations were not stretched. Against that backdrop, a strong semiconductor export number would have room to force a sharp reversal — sellers have to come back in — which is precisely the setup for the kind of chip-led reversal pattern sketched in the introduction. Whether it plays out that way on any given release day depends on the print itself, but the positioning math frames the risk-reward. Had foreigners been buying 70,000-plus eok won into the release instead, the same beat would have been at least partly pre-positioned, and the risk of a sell-the-news reaction correspondingly higher.
Two domestic liquidity gauges help judge whether a print-driven move has follow-through fuel. Tujaja yetakgeum (investor deposits — customer cash sitting in brokerage accounts) stood at roughly 104.2 trillion won as of 2026-08-07, and sinyong yungja (margin loans — money borrowed to buy stocks) at roughly 29.2 trillion won on the same date. Rising deposits alongside restrained margin debt suggest dry powder without leverage froth; the reverse combination means a rally is being chased with borrowed money and is more fragile.
Common failure modes
- Extrapolating one 10-day period. Ten days is a small sample. Shipment timing, a single large ship delivery, or quarter-end front-loading can distort the window. Treat the 10-day print as a direction check and wait for the 20-day print before updating full-month views with confidence.
- Ignoring the ship line. Vessel deliveries are enormous and lumpy — one delivered LNG carrier can move the headline growth rate. If the headline surprises but semiconductors did not, check whether ships explain the gap before drawing macro conclusions.
- Confusing price and volume. A jump in semiconductor export value during a memory price spike says less about unit demand than the same jump during flat prices. Cross-check with whatever memory pricing commentary you follow before calling a demand inflection.
- Forgetting the import side. Falling chip-equipment imports while chip exports still look fine is an early warning that producers are cutting capex — often before the export line rolls over.
- Holiday-shifted comparisons. Any January/February or September/October print should be read only on a daily-average basis, and ideally with the two affected months averaged together.
FAQ
Where and when exactly is the data released?
The Korea Customs Service publishes the figures on or around the 1st, 11th and 21st of each month, in the morning Seoul time, moving to the next business day when the date falls on a holiday. Wire services carry the headline numbers within minutes, and Korean financial portals post the breakdown tables the same morning.
Does the 10-day print reliably predict the full month?
Directionally it is useful; numerically it is noisy. Working-day composition, shipment timing and lumpy categories like ships mean the 10-day growth rate regularly differs from the eventual monthly figure by several percentage points. The 20-day print is a much tighter predictor. Use the 10-day release to test whether a trend is intact, not to forecast a precise monthly number.
Why do the headline YoY and daily-average YoY sometimes point in opposite directions?
Because the same calendar window can contain a different number of working days than a year earlier, especially around Seollal and Chuseok. The headline compares total values; the daily average corrects for the calendar. When they conflict, the daily average is the economically meaningful one, and markets generally converge on it within the session.
Sources
- KRX data portal — data.krx.co.kr (index levels, investor-type trading flows)
- Naver Finance — finance.naver.com (daily foreign net buying by market)
- KOFIA statistics — freesis.kofia.or.kr (investor deposits, margin loan balances)
- Bank of Korea ECOS — ecos.bok.or.kr (trade, FX and macro time series)
Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.