Korea exported $98.25 billion in August 2026, up 68.7% from a year earlier. The daily-average export growth rate for the same month was 72.5%. Two growth rates drawn from the identical month's trade data sit 3.8 percentage points apart. What separated them was not demand or the exchange rate but the calendar — August 2026 had half a working day fewer than August 2025.
Read the headline number alone and that calendar effect gets mistaken for a change in the business cycle. This is precisely why Korea's Ministry of Trade, Industry and Energy publishes the headline and daily-average figures side by side every month. The arithmetic is one division and one multiplication, and the same correction applies cleanly to individual product categories.
How daily-average exports are calculated
Daily-average exports = monthly export value ÷ working days. According to the ministry's August 2026 trade report, August exports were $98.25 billion across 22.0 working days. 98.25 ÷ 22.0 = 4.466, matching the published daily average of $4.47 billion.
August 2025 had 22.5 working days. A 68.7% increase implies that August 2025 exports were 98.25 ÷ 1.687 = $58.24 billion. Divide by 22.5 and you get $2.59 billion — exactly the prior-year daily average as published. Comparing the two daily averages, 4.47 ÷ 2.59 = 1.725, or +72.5%. Working backwards from the published figures lands on the same number.
Working-day counts land on half-day increments like 22.0 and 22.5 because Saturdays count as 0.5. Monday through Friday count as one day, Saturday as half, Sunday and public holidays as zero. So the count for the same calendar month shifts year to year depending on where the lunar holidays fall and how many substitute holidays are designated. Even August, with Liberation Day as its only public holiday, moves by half a day depending on which weekday that date lands on.
The reason for publishing a daily average at all is simple. Monthly export value scales mechanically with how many days customs clearance was possible. Without fixing the day count, year-on-year comparison breaks down several times a year.
Where the 3.8pp between 68.7% and 72.5% comes from
Written as a formula: daily-average growth = (1 + headline growth) × (prior-year working days ÷ current working days) − 1. August's adjustment factor is 22.5 ÷ 22.0 = 1.0227. Multiply the headline multiple of 1.687 by 1.0227 and you get 1.7253 — +72.5%, to the decimal. It matches the figure reported by Money Today.
The adjustment factor applies identically regardless of product. Feeding August's published category growth rates through 1.0227 produces the table below. The third column is as published; the fourth is calculated here by applying the factor.
| Category | Aug exports ($bn) | Headline growth (%) | Working-day adjusted (%, calculated) | Gap (pp) |
|---|---|---|---|---|
| Semiconductors | 46.7 | +209 | +216.0 | +7.0 |
| Computers | 6.2 | +419 | +430.8 | +11.8 |
| Petroleum products | 6.8 | +65 | +68.8 | +3.8 |
| Petrochemicals | 3.9 | +12 | +14.5 | +2.5 |
| Automobiles | 3.9 | −30 | −28.4 | +1.6 |
| Ships | 1.7 | −46 | −44.8 | +1.2 |
| Total | 98.25 | +68.7 | +72.5 | +3.8 |
Notice that the size of the gap scales with the growth rate itself. Semiconductors, already up 209%, gain another 7.0pp when adjusted; petrochemicals, up 12%, gain only 2.5pp. The working-day correction multiplies the growth multiple rather than adding a fixed amount, so the larger the increase, the wider the adjusted gap.
For declining categories the direction reverses. Automobiles move from −30% to −28.4%, ships from −46% to −44.8%. Normalised to the same number of days, the actual contraction was smaller than the headline suggested. August semiconductor exports of $46.7 billion cleared $40 billion for a third straight month, but volume indicators like this only mean something read alongside price indicators — the framework is laid out in this piece on the gap between DRAM contract and spot prices.
One working day swings the growth rate by 4.5pp
Monthly working days typically range between 20 and 23. Take 22 as the base: even in a month where actual daily-average exports were flat year on year, one extra working day (23 ÷ 22 = 1.045) publishes as a 4.5% increase, and one fewer (21 ÷ 22 = 0.955) as a 4.5% decrease. That is ±4.5% generated by the calendar alone in a month where nothing happened.
The magnitude is not trivial. Years where Seollal falls in January versus February show double-digit swings in the same month's headline growth. That is why monthly export releases carry a "holiday effect" caveat. Conversely, when working days match the prior year exactly, the adjustment factor is 1 and the two growth rates become the same number. A month where both figures agree is a month where the calendar did not intervene.
Daily-average adjustment is not a cure-all, though. Working days count business days on which customs clearance happens, not days a factory ran. Product lines that run 24 hours — semiconductors above all — keep producing through holidays and merely push clearance to the next business day. For those categories the adjustment can overstate the underlying gain. The correction assumes shipments scale with business days, and where that assumption is weak the adjusted figure is only a reference point.
The opposite extreme exists too. Ship exports are booked when contractual delivery lands, so they cluster in particular months. That is why adjusting August's 46% decline to −44.8% changes nothing about the interpretation. Working-day correction earns its keep in consumer goods and components, where shipments go out evenly day by day.
The headline growth rate shows you the market and the calendar mixed together; the daily-average rate strips the calendar out. The gap between them is the calendar's share of that month.
What to watch
- On the first of each month, read the ministry's headline and daily-average growth rates together — the gap is the size of that month's working-day effect.
- Check the prior-year and current working-day counts printed in the release. The adjustment factor is nothing more than prior ÷ current.
- In years where holidays fall in a different month than the year before, do not judge January–February or September–October data on headline growth alone.
- The same factor applies to category-level growth — multiply (1 + growth) by the factor before comparing published figures against each other.
- The trade balance is a difference of totals, so it is not subject to daily-average adjustment. Just note that a month with fewer working days books a smaller surplus outright. August's surplus was $34.75 billion; the January–August cumulative surplus was $202.5 billion.
- If daily-average growth keeps printing below headline growth for several months, working days are expanding — check whether the strong headline is really just extra days.