Chuseok falls on 25 September this year. Every August the same sentence starts circulating in Korea: “the holiday shop costs more than last year.” So we counted. We pulled the whole farm produce price feed that the Korea Agro-Fisheries & Food Trade Corporation (aT) publishes through the national open data portal, kept only the items that arrive with a year-ago figure attached, and compared them one by one.
1. Of 74 retail items, 54 fell. On the survey dated 11 August 2026 the median change was −7.2% against a year earlier. Twenty items rose.
2. But an average would lie. Inside the same fruit category, Singo pear is +24.9% while Tsugaru apple is −29.1% — 54 percentage points apart.
3. Retail and wholesale disagree. The wholesale-dealer selling price barely moved, at a median of −0.8%. Matching the same item, variety and grade gives 62 pairs, and retail fell further in 44 of them.
Far more items fell than rose
The feed returned 452 rows. Retail and wholesale prices are mixed together, one row per item, variety and grade. After removing rows whose survey had stopped years ago (explained further down) and keeping only rows that carry a year-ago figure, 74 retail items and 164 wholesale items remain.
Across those 74 retail items the median change was −7.2%. Twenty rose, 54 fell. We use the median rather than the mean for a reason that shows up in the next section: one tail is long enough to drag a mean around by itself.
Split by category, vegetables and fruit carried the fall. Grains barely moved.
| Category (retail) | Items | Median | Up / down |
|---|---|---|---|
| Vegetables | 39 | −8.7% | 9 / 30 |
| Fruit | 15 | −11.1% | 5 / 10 |
| Grains | 8 | −2.9% | 3 / 5 |
| Special crops | 6 | −5.3% | 1 / 5 |
| Seafood | 6 | −4.7% | 2 / 4 |
| Total | 74 | −7.2% | 20 / 54 |
One thing to state before going further: this feed has almost no livestock in it. Seven rows are classified as livestock, and the newest of them was surveyed on 1 April 2022. So the numbers here describe a produce basket, not the whole grocery basket. Beef and eggs are not in this data.
On the same table, some things rose and some fell
Here are the items that most often reach a Korean holiday table. Pears rose and apples fell — same category, opposite directions.
| Item (retail) | Change | Price now | Survey unit |
|---|---|---|---|
| Pear, Singo | +24.9% | 44,516 won | 10 fruit |
| Pollack, frozen | +12.5% | 3,877 won | 1 fish |
| Rice | +4.8% | 61,350 won | 20kg |
| Sesame, white | −3.6% | 17,800 won | 500g |
| Carrot, unwashed | −4.9% | 3,443 won | 1kg |
| Spinach | −15.6% | 2,016 won | 100g |
| Apple, Fuji | −20.6% | 25,496 won | 10 fruit |
| Radish, highland | −25.6% | 1,955 won | 1 piece |
| Apple, Tsugaru | −29.1% | 19,517 won | 10 fruit |
| Napa cabbage, summer | −32.4% | 4,492 won | 1 head |
Do not read the price column downwards. The survey unit changes from row to row — rice is priced per 20kg, spinach per 100g. The only column that compares across rows is the change, because each figure there compares one item against itself a year earlier in the same unit.
So “it is expensive this year” and “it is cheap this year” are both true and both wrong. A household that puts apples on the table spends noticeably less than last year; a household that puts pears on it spends more. What is in the basket mattered far more than the overall price level.
Retail fell; wholesale did not
The feed splits every row into retail or wholesale-dealer selling price. Counted separately, the two give different answers.
The median across 164 wholesale items is −0.8%: 76 rose and 88 fell, close to an even split. Retail was 20 against 54.
That could be an artefact of the two sets containing different items, so we matched them — same item, same variety, same grade. That produced 62 pairs, and within those pairs the retail median was −8.2% against a wholesale median of −1.4%. In 44 of the 62 pairs, retail fell further than wholesale.
| Item (standard grade) | Retail | Wholesale |
|---|---|---|
| Zucchini | +12.0% 1 pc | +118.3% 10kg |
| Green chilli | −18.1% 100g | +50.9% 10kg |
| Korean zucchini | +15.2% 1 pc | +49.3% 20 pc |
| Korean melon | −27.7% 10 pc | +5.0% 10kg |
| Lettuce, green | −7.5% 100g | +22.1% 4kg |
| Ginger, domestic | +14.2% 1kg | −6.2% 10kg |
| Bell pepper, green | −3.2% 100g | −27.1% 10kg |
| Paprika | −24.6% 1 pc | −51.8% 5kg |
The grey text is each side’s own survey unit. Retail is surveyed by the piece or per 100g; the wholesale-dealer price is usually a 10kg or 5kg case. Counting all 62 pairs, only six have the same unit on both sides. That is why this table carries no amounts at all — the only comparable thing is each side’s change against its own year-ago price in its own unit.
We are not going to draw a conclusion about distribution margins from this. Retail and wholesale are surveyed in different places, with different grade mixes and different selling units. What this does establish is narrower: look at one side only and you reach the opposite conclusion. It is also why Korean news can carry “wholesale prices rose” and “retail prices fell” in the same week. The same feed is embedded on our live data page, where you can switch between the two yourself.
The official price index points the same way
What we counted are surveyed prices, not a price index, so we checked our answer against a different institution using a different method. The July 2026 Consumer Price Trends briefing, released by Korea’s National Data Office on 4 August 2026, says:
The fresh food index fell 0.8% from the previous month and 2.3% from a year earlier. […] Agricultural, livestock and fishery products fell 0.8% from the previous month and rose 0.9% from a year earlier […]
Headline consumer prices that month were up 2.8% year on year. So the general price level rose while fresh food alone fell — and the broader grouping that folds livestock and processed goods back in was up 0.9%. Our figure of −7.2%, computed on a basket with no livestock in it, points the same way. Different samples and different methods have no reason to produce the same number, but both say fresh produce got cheaper.
Do not trust the dates in this feed at face value
The endpoint is named “latest-date wholesale and retail prices”. That does not mean one recent survey date for everything. What actually arrives is each item’s own last survey date. Where a survey has been discontinued, a price from years ago rides in looking like today’s.
| Survey year | Rows | Share |
|---|---|---|
| 2026 | 281 | 62.2% |
| 2025 | 28 | 6.2% |
| 2020–2024 | 33 | 7.3% |
| 2010–2019 | 81 | 17.9% |
| 2009 or earlier | 29 | 6.4% |
| Total | 452 | 100% |
The oldest row was surveyed on 22 August 1997 — a peach variety. Only 234 rows were surveyed on 11 August 2026 itself. Every figure in this article is therefore computed on the 247 rows within 30 days of the newest survey, with 205 rows dropped. We write that down rather than passing over it quietly, because leaving those 205 rows in would put a 1997 price on screen as today’s.
Why the amounts cannot be compared to each other
The feed also supplies a per-kilogram converted price alongside the surveyed one. It looks convenient, and it has a trap: the conversion only actually happens for items surveyed in kg or g. Items surveyed by the piece, head, fish or pair arrive with the surveyed price copied straight into the converted field.
Mix the two and “61,350 won per 20kg” turns into “3,068 won per 20kg”, because rice converts to a per-kilogram figure. In the other direction, beef at 100g is surveyed at 3,472 won but converts to 34,717. The values are right; it is the label that stops matching. So both here and in the live widget we print the surveyed value as it arrives, with its survey unit next to it.
What that means for the holiday shop
Only what the numbers support. With fruit, the variety decides it — on these figures apples are far kinder than pears this year. Napa cabbage, radish and spinach are all clearly cheaper than a year ago. Rice and pollack rose.
All of that is the 11 August survey, though. Prices move again as the holiday peak arrives, so it is worth checking the day’s figure on the live data page before you shop. How long the food keeps once you have bought it is covered in sell-by versus use-by dates.
A few other places money moves over the holiday: expressway toll exemption covers only three of the four holiday days, and if you work through it you can check what holiday work pay comes to. What actually sends people to hospital over Chuseok is in the holiday accident figures.
Sources used
- Korea Agro-Fisheries & Food Trade Corporation — KAMIS open API — “Latest-date wholesale and retail price enquiry” on data.go.kr. Every change, median and item count here was computed directly from the 452 rows this API returned on 12 August 2026. The portal lists the provider as aT, the cost as free and the licence as unrestricted. Its description reads “average wholesale-dealer selling price and retail price for one day, one week, one month and one year before the query date” — which is why this article says wholesale-dealer selling price rather than simply wholesale.
- National Data Office — July 2026 Consumer Price Trends — the release itself (Price Trends Division, posted 4 August 2026). The sentences quoted above come verbatim from the briefing transcript of the same release. The fresh food index −2.3%, agricultural, livestock and fishery products +0.9% and headline CPI +2.8%, all year on year, are quoted from it.
- The holiday dates — taken from the reading of the Public Holidays Regulation and the Toll Road Act Decree we did for the toll exemption piece.
What we could not check
- Chestnut, jujube, dried persimmon, bellflower root and bracken were not in this feed. None of these ritual-table staples appeared in any of the 452 rows, so we could not measure them — whether another survey covers them, we did not check.
- Beef, pork and eggs are excluded. All seven livestock rows were surveyed before April 2022, so this feed cannot answer for them. Livestock prices are published separately by the Korea Institute for Animal Products Quality Evaluation.
- Regional differences were not examined. This endpoint returns no regional breakdown, so we treated every figure as national in character.
Written as at August 2026. All percentage changes, medians and item counts were computed directly from the 452 rows of the 11 August 2026 survey, using the 247 rows that remained after 205 stale rows were dropped. Because survey units differ by item, amounts were never compared across items. Livestock and the ritual-table forest products could not be checked. Other everyday calculations are in the calculator collection.
