Working days in Korea's trade data — how to subtract two days from a 102% export jump
In South Korea's trade statistics, working days count how many business days fall in a period: by convention weekdays count as one, Saturdays as half, and Sundays and public holidays as zero. Lunar holidays such as Chuseok and Seollal move around the calendar, so the same month can have two or more extra or missing working days from one year to the next, inflating or shrinking headline growth. Dividing exports by working days gives daily average exports, the cleaner measure. From October 1 to 10, 2026, exports rose 102.0%, but working days were 5.5 against 3.5 a year earlier (Chuseok), so daily average exports rose 28.5%.
The three lines
- Definition — business days in the period; by convention weekdays 1, Saturdays 0.5, Sundays and holidays 0
- Why it matters — moving lunar holidays shift working days by two or more, swinging headline growth
- Read order — daily average growth before headline growth. October 1-10, 2026: headline +102.0%, daily average +28.5%
Key questions
- What are working days in Korea export data
- **The number of business days in the period, used to compare exports fairly.** | Day | Counts as | |---|---| | Weekday | 1 | | Saturday | 0.5 | | Sunday | 0 | | Public / substitute holiday | 0 | | Example: Oct. 1-10, 2026 | 5 weekdays + 1 Saturday (0.5) = 5.5 |
- How to calculate working days
- **Add 1 per weekday, 0.5 per Saturday, 0 for Sundays and holidays. For October 1-10, 2026 that gives exactly the official 5.5.** | Date | Day | Type | Count | |---|---|---|---| | Oct. 1-2 | Thu-Fri | Weekdays | 2 | | Oct. 3 | Sat | National Foundation Day | 0 | | Oct. 4 | Sun | — | 0 | | Oct. 5 | Mon | Substitute holiday | 0 | | Oct. 6-8 | Tue-Thu | Weekdays | 3 | | Oct. 9 | Fri | Hangul Day | 0 | | Oct. 10 | Sat | Saturday | 0.5 | | Total | — | — | 5.5 |
- Daily average exports calculation
- **Divide total exports by working days. It removes the calendar effect.** | Item | Oct. 1-10, 2025 | Oct. 1-10, 2026 | |---|---|---| | Exports | ~$12.9B | $26.1B | | Working days | 3.5 | 5.5 | | Daily average | ~$3.69B | $4.75B | | Growth | Headline +102.0% | Daily average +28.5% |
The first number to check in a Korean trade release is not growth — it's working days. Exports for October 1-10, 2026 rose 102.0% from a year earlier. Doubling sounds dramatic, but last year that ten-day window held just 3.5 working days, and this year 5.5. With 57% more working days, some of the jump was simply arithmetic. Working days are the tool for stripping that out.
1. How working days are counted
| Day | Counted as | Why |
|---|---|---|
| Weekday | 1 | Factories, ports and customs fully operating |
| Saturday | 0.5 | Some plants and logistics operating |
| Sunday | 0 | — |
| Public / substitute holiday | 0 | Lunar holidays, national days |
| Around holidays | Per date type | Company-specific closures handled separately |
Apply this to October 1-10, 2026 and you hit the official figure exactly: October 1-2 (two weekdays), zero for National Foundation Day (Saturday the 3rd), Sunday the 4th and the substitute holiday on Monday the 5th, three weekdays from the 6th to 8th, zero for Hangul Day on Friday the 9th, and 0.5 for Saturday the 10th — 5.5 days. In 2025 the Chuseok holidays ran from National Foundation Day on October 3 through Hangul Day on October 9, leaving only three weekdays (1, 2 and 10) plus Saturday the 4th: 3.5 days.
2. Reading the same data two ways
| Period | Headline growth | Working days | Daily average growth | Takeaway |
|---|---|---|---|---|
| Oct. 1-10, 2026 | +102.0% | 3.5 → 5.5 | +28.5% | Calendar inflated the headline; daily average still a record |
| Jan. 1-10, 2026 | -2.3% | 7.5 → 7.0 | +4.7% | Half a day fewer made exports look weaker |
| January 2025 | Decline | 24 → 20 | +7.7% | Seollal took four days out |
| Sept. 1-20, 2026 | +78% | One fewer | Higher still | Grew despite fewer days — genuine strength |
A decline can mean fewer days or weaker demand; only the daily average tells you which. Likewise a surge can mean more days or more sales. When exports soar despite fewer working days — as in September 1-20 — that is the strongest signal.
3. How it differs from a base effect
| Item | Working-day effect | Base effect |
|---|---|---|
| Cause | Different number of business days | The comparison period itself was abnormally low or high |
| Typical case | Which month Chuseok or Seollal lands in | Last year's slump, strike or accident |
| Fix | Divide by working days | Compare with two years ago, look at trend |
| Overlap | Often — a holiday-depressed base triggers both | — |
The working-day effect can be seen as one kind of base effect: last October's holiday-shortened period produced a low base that inflated this year's growth.
4. A checklist for trade headlines
| Step | Check | Why |
|---|---|---|
| 1 | Working days, this year vs. last | A gap of 0.5 days or more means don't trust headline growth |
| 2 | Daily average growth | Calendar-free underlying figure |
| 3 | Whether items beat the working-day effect | In October 1-10, 2026 chips (+170.9%) did; steel (+49.4%) did not beat the 57% rise in days |
| 4 | Preliminary or final | 10- and 20-day data are preliminary; monthly releases are the benchmark |
| 5 | Imports and balance | If energy imports rise faster, the surplus shrinks |
5. Frequently asked questions
| Question | Answer |
|---|---|
| Why is Saturday half a day? | By convention, since some plants, logistics and customs operate on Saturdays even under a five-day week |
| Do monthly releases use it? | Yes; monthly data also report working days and daily averages |
| If working days match, can I skip the daily average? | Headline and daily growth converge, though pre-holiday "push" shipments can still distort |
| Do other countries do this? | Practices vary; many publish seasonally or working-day adjusted series |
6. What remains unconfirmed
- Official formula: The Saturday-half-day convention fits this release exactly, but we didn't find the government's methodology document.
- Company shutdowns: How voluntary closures around holidays are reflected is unconfirmed.
- Full October: Monthly working days and daily averages arrive with the early-November release.
Sources
- Gukje News — October 1-10 exports $26.1 billion (5.5 working days, daily average +28.5%)
- eToday — Early-January exports fall 2.3% on fewer working days
- Aju News — September exports down 8.2% on fewer working days; daily average record
- Metro Seoul — Seollal holidays cut January working days (February 2025)