SeoulMarkets · rates

Korea lists bonds on three boards

Latest session 2026-09-09 · 22 sessions in window · updated each trading day

Almost every published Korean bond yield comes from the government inter-dealer board, which carried 10 listings on the latest session. Beside it sit the general board and a near-fixed 40-listing small-lot board that Korean households actually buy from — and nobody counts it in English. This is the compass, free. The full daily history and API are what we licence.

BoardListingsMedian yieldOver government
Government (inter-dealer)103.878%
Small-lot404.477%0.599pp
General2714.645%0.767pp

The gap is not a constant

Board over governmentNarrowestWidestLatest
General0.412pp0.950pp0.767pp
Small-lot0.426pp0.774pp0.599pp

The government board is not a fixed reference either — its own median ran 3.630% to 4.011% inside the same window, so the premium widens and narrows for reasons on both sides of it.

Sorted by who issued the paper

The board a bond trades on is a different question from who owes the money. Issuer type is read out of the listing name, and names that do not announce it stay unclassified rather than being folded into a tidier bucket.

Issuer typeListingsMedian yield
Financial1005.049%
Unclassified964.731%
Local government324.468%
Public corporation334.406%
Treasury404.386%
Housing bonds204.043%

How this is counted, and what is left out

  • Every figure is a median, never a mean. One general-board listing settled at 117.082% on 9 September; a mean would have carried it into the headline.
  • Yields outside 0.1-20% are counted and reported, not deleted11 rows of 321 on 2026-09-09. They are what the source published, and a record that quietly drops them is not a record.
  • Opening, high and low yield fields are not used anywhere on this page. 22 of 3,547 rows carried a negative high yield, and the widest intraday range those fields imply is 23,602 percentage points, which is not a market move.
  • Residual maturity and classification arrive filled on 2.8% of rows, so no maturity bucket appears here. Maturity is recoverable from listing names and is used on our yield-curve work instead.
  • A board with fewer than three listings on a day is shown as unmeasured, not given a median.
  • This series had a hole. Collection stopped after 24 August and went unnoticed for 18 days because the dataset sat outside our missing-day check. The 12 absent sessions were refetched from source on 11 September and the check now covers it.

Source: Financial Services Commission (Republic of Korea), Bond Securities Information Service. Medians computed by SeoulMarkets. This is not investment advice.