A snapshot from Encar (encar.com), Korea's largest used-car marketplace — 223,000 live listings — pulled 2026-07-13. Public listing data only; no personal or dealer data.
Korea is one of the world's larger used-car export sources, and if you're on the buying end of that trade, the question that matters isn't "what's for sale" — it's "where is listing activity concentrated right now?"
So I took the simplest possible cut: the 800 most-recently-listed cars on Encar, and compared their shape to the whole 223K-listing market. One honesty note up front: "most-recently-listed" includes re-lists and bumped ads, not only cars newly entering the market — so everything below is about listing activity, not proven new supply. The pull took one run of encar-scraper, an Apify actor I maintain that turns Encar's listing feed into clean JSON. If fresh listing activity looked like the standing market, the two should line up.
They didn't.
One province is running 4.5× above its weight
If you work this trade you already know the southeast runs Korea's used-car exports. What I hadn't seen before is that quantified against standing stock. Break the newest 800 down by region and compare each region's share of fresh listings to its share of total inventory:
| Region | Share of newest listings | Share of total market | Index |
|---|---|---|---|
| 경남 South Gyeongsang | 13.0% | 2.9% | 4.5× |
| 대구 Daegu | 16.5% | 9.0% | 1.8× |
| 인천 Incheon | 8.8% | 6.1% | 1.5× |
| 서울 Seoul | 11.3% | 10.3% | 1.1× |
| 부산 Busan | 8.0% | 9.7% | 0.8× |
| 경기 Gyeonggi | 31.4% | 42.7% | 0.7× |
South Gyeongsang holds under 3% of Korea's used-car inventory — but accounted for 13% of the freshest listings in this snapshot. Daegu, next door, ran at nearly double its weight. Together, Daegu + South Gyeongsang = 29.5% of fresh listing activity from 11.9% of the standing stock — a 2.5× concentration.
Meanwhile Gyeonggi — the Seoul-belt province that holds 43% of all inventory — under-indexed (0.7×). The big pool is sitting still; the southeast corner is the one churning.
That corner is also where Korea's export machine lives — Busan port and the Daegu auto-export complex. Two caveats before reading it that way: this is one cross-sectional snapshot, so it can't prove causation, and a handful of high-volume southeastern dealers re-bumping their inventory could account for part of the anomaly. Still, "which region is actively churning listings" is a signal worth watching if you source there — and it's not where a population map would point you.
What's over-represented in fresh listings — and what actually exports
The single most-listed model across the entire fresh feed was the BMW 5-Series (G30) — 19 units, over-indexing at 1.13× (Mercedes, for what it's worth, under-indexed at 0.66× — this is a BMW story, not a "German cars" story). Genesis over-indexed at 1.3×, with 47 units of GV70, GV80, and G80 inside the top 15.
Top fresh models:
- BMW 5-Series (G30) — 19
- Genesis GV70 — 18
- Hyundai Grandeur IG — 18
- Kia Ray — 16
- Hyundai Avante AD — 15
- Genesis GV80 — 15
- Kia Carnival (4th gen) — 15
To be clear, this list is a mix: the 5-Series, Genesis SUVs, and Grandeur are export-relevant metal, while the Ray, Avante, and QM6 are domestic-turnover cars that mostly stay home. The signal isn't "everything fresh is export-bound" — it's that the export-relevant segment (premium sedans, Genesis SUVs) is over-represented in fresh listing activity rather than diluted by it. Median fresh asking price: ₩20.9M (~$15,100).
The glut that wasn't in the data
Every few months there's a headline about Korea's used-EV glut flooding the second-hand market. In this snapshot, it didn't show up.
Pure electric cars were 1.9% of the freshest 800 listings — lower than their 3.1% share of standing inventory (a 0.6× index). Whatever is churning through Korea's used market right now, it is overwhelmingly gasoline (60%) and diesel (24%). One snapshot doesn't debunk a narrative, but if EVs were pouring in, this is where it would show — and it didn't.
The age band lines up with export demand
Fresh listings cluster at 2019–2023 model years — 3 to 7 years old. I won't claim intent behind that; a healthy used market naturally churns cars at that age. But it happens to be the band most export destinations' age caps target, which makes the fresh feed unusually usable for sourcing: for buyers, the overlap matters more than the reason.
Reproduce this yourself (~5 minutes, free)
Everything above is one actor run away. On encar-scraper, this input pulls the same fresh-listings sample:
{
"carType": "all",
"sort": "newest",
"maxItems": 800
}
Each record comes back as structured JSON — make, model, year, price_krw, mileage_km, fuel, region, url, image_url — so the region/fuel/price group-bys in this post are a few lines of pandas (or a pivot table; CSV/Excel export works too). Apify's free plan includes $5 of monthly credit, and at $2 per 1,000 results that covers this entire analysis with room to spare — no credit card needed.
Want the second half of the market? Korea's #2 marketplace runs on the same contract: kbchachacha-scraper returns identically-shaped records (~168k more listings), so the two feeds concat into one dataframe cleanly. If your stack talks MCP, both actors also work as AI-agent tools out of the box.
Why I'm posting this
I run a suite of 24 Korean-platform scrapers — Korean sites are largely invisible to Western data tools, and I keep finding stories like this one sitting in plain sight. Everything here is public, non-authenticated listing data: the actor reads Encar's own listing feed and never touches logins, private endpoints, or personal information.
The raw listings aren't the interesting part — anyone can browse those a page at a time. But structured, one afternoon's pull surfaces a regional churn anomaly, a premium-listing skew, and an EV glut that failed to appear — patterns that are hard to see at a glance even if you live in this market.
Next on my list is turning this into a live export-sourcing feed: newest listings by make and region, with destination age-cap eligibility computed per car from model year. If you source cars out of Korea, tell me in the comments which filters would actually matter to you — that's what gets built first.
Method: 800 newest-listed cars (Encar public list API, sort = most-recently-listed) vs whole-market count baselines, 2026-07-13, from a Korean IP. Single cross-sectional snapshot, not a time series — every comparison here is "fresh listings vs standing stock," never "this month vs last." "Fresh" includes re-listed ads. Public listing attributes only; no dealer or personal data collected.
Top comments (0)