Korea Dividend Season: December Record Dates & the Ex-Dividend Cliff

Korea compresses most of its dividend calendar into the last trading days of December: the shareholder register is fixed at year-end, yet the actual dividend amount is typically not confirmed until the annual general meeting in February or March. That single quirk — rights first, amount later — explains almost everything strange about Korean year-end flows: the arbitrage buying into the record date, the sharp one-day drop on the ex-dividend date, and why preferred shares suddenly become popular in November. If you understand the record-date mechanics and watch the foreign net-flow series, December stops looking random.

Why Everything Happens in December

The vast majority of listed Korean companies use the calendar year as their fiscal year. Historically, their articles of incorporation set the dividend record date (gijunil, the date on which the shareholder register is fixed) as the final day of December. The Korea Exchange (KRX) does not trade on the last business day of the year, so the final session falls just before year-end, and Korea settles equities on a T+2 basis.

Chaining those three facts together gives you the structure of every Korean dividend season:

  1. To be on the register at the record date, your purchase must settle by then.
  2. With T+2 settlement, the last day you can buy and still receive the dividend (the last cum-dividend day) falls two trading sessions before the register is fixed — in practice, the final week of December.
  3. The next session is the baedangnak-il (ex-dividend date): anyone buying from that day on gets no dividend, and the price typically gaps down to reflect it.

The exact calendar dates shift slightly each year with weekends and holidays, so always confirm the current year’s cum-dividend deadline on the KRX data portal or your broker’s notice rather than assuming a fixed date.

The Blind-Dividend Problem — and the Slow Fix

Here is the part that surprises foreign investors most: under the traditional calendar, you commit capital at the December record date without knowing the dividend amount. The board proposes the payout in late winter and shareholders approve it at the AGM, months after rights were fixed. Practitioners call this the blind-dividend problem — you are buying a coupon of unknown size.

Since roughly 2023, Korean regulators have encouraged companies to amend their articles so the record date comes after the dividend amount is announced (“amount first, record date later”). Adoption has been gradual rather than universal, which means Korea currently runs a split calendar: reformed companies set record dates in the first quarter after disclosing the payout, while legacy companies still use the year-end register. Before trading any dividend angle on a specific name, check which regime it follows — the KRX disclosure system and Naver Finance both show record dates per company. Assuming the old December date for a reformed company is now one of the most common retail mistakes in this trade.

The Ex-Dividend Cliff: No Official Markdown

A structural detail that matters for anyone reading Korean charts: KRX adjusts the reference price on the ex-date only for stock dividends. For cash dividends there is no official base-price adjustment — the market simply reprices the share on its own on the ex-dividend morning. The result is the familiar “cliff”: high-yield names often gap down by something close to the expected dividend, sometimes more, sometimes less, depending on sentiment and positioning.

Two practical consequences:

  • Screen distortion. A high-dividend stock showing a sharp single-day loss in late December may have lost nothing economically once the accrued dividend is counted. Year-end performance screens that ignore this will mislead you.
  • The naive capture trade usually fails. Buying on the last cum-dividend day purely to collect the payout means absorbing the cliff plus withholding tax (roughly 15.4% for Korean resident individuals; treaty rates vary for foreigners). Unless the ex-date drop is smaller than the after-tax dividend — which you cannot know in advance, especially for blind-dividend names — the expected edge is near zero. The investors who do profit around the record date are running hedged structures, not outright buys.

Who Actually Arbitrages Dividend Season

The December flows that move the tape come mainly from institutions and foreign investors running relative-value structures:

  • Index basis trades. KOSPI 200 futures do not receive dividends, so as the December record date approaches, expected dividends get priced into the basis and futures tend to trade cheap relative to the cash basket. Arbitrage desks buy the stock basket and sell futures, collecting the dividend as the carry. These program flows show up as concentrated buying in large-cap index names in the final weeks of the year, then unwind after the ex-date.
  • Dividend-capture rotation. Pension funds and other institutions rotate toward high-payout sectors — historically banks, insurers, telecoms and utilities — in November and December, then let positions normalize in the new year.
  • The preferred-share switch. Korean preferred shares (useonju — non-voting shares that typically pay a slightly higher dividend than the common shares, botongju) have historically traded at steep discounts to the commons, often in the region of 30–50% for large issuers. A cheaper entry price on a similar or higher dividend means a materially higher yield, so the preferred discount tends to narrow into the record date and widen again after the ex-date. Watching a common/preferred pair through December is one of the cleanest ways to see dividend season in the data.

Worked Example: Reading the Foreign Net-Flow Series

Dividend-season flows ultimately show up in the daily investor-flow tables, so let’s read a real one. Below are foreign investors’ daily net purchases on KOSPI for nine completed sessions in October 2026, in eok won (1 eok = KRW 100 million; 10,000 eok = KRW 1 trillion):

Session (2026) Foreign net flow (eok won)
Oct 12 -20,103
Oct 13 -45,789
Oct 14 -11,614
Oct 15 -37,637
Oct 16 +3,637
Oct 19 +947
Oct 20 -4,884
Oct 21 +6,462
Oct 22 +505

Step by step:

  1. Convert the unit. The Oct 13 print of -45,789 eok won is -45,789 × KRW 100 million = roughly KRW 4.6 trillion of net selling in a single session — a genuinely heavy day. At the won’s late-October 2026 level of around 1,326 per dollar (1,325.59 at the 2026-10-22 close), that is on the order of USD 3.5 billion.
  2. Read the sign pattern, not one day. Four consecutive heavy selling sessions (Oct 12–15) are a regime; the small positive prints that follow (+3,637, +947, +6,462, +505) suggest the selling wave exhausted rather than reversed with conviction.
  3. Cumulate. Summing all nine sessions gives roughly -108,000 eok won, i.e., about KRW 10.8 trillion of net foreign selling over two weeks. The cumulative figure is the signal; any single day is noise.

Why this matters for dividend season: if legacy-calendar dividend capture is happening, you should see this same series turn into a sustained positive run through late November and December, concentrated in index heavyweights, followed by give-back after the ex-dividend date. If December arrives and the foreign series stays negative, the arbitrage bid is absent that year — useful information regardless of your view. For context on the domestic side, retail liquidity was substantial as of October 2026: customer deposits at brokerages stood at about KRW 89.6 trillion and margin lending at about KRW 30.3 trillion (both as of 2026-10-21, KOFIA data), against a KOSPI that closed at 5,533.77 on 2026-10-22.

The Shift Toward Quarterly Dividends

The single-date December structure is slowly dissolving. Samsung Electronics has paid quarterly dividends for years, and a growing group of large caps has followed with semiannual or quarterly schedules. The Corporate Value-up Program launched in 2024 added regulatory pressure for clearer, policy-based payouts — published payout ratios, buyback plans, and record dates set after amounts are known.

For flow watchers, the implication is that the December concentration should keep fading: more record dates scattered across the year means smaller single-date arbitrage, shallower cliffs, and less violent preferred-discount cycles. The mechanics in this guide still apply to each individual record date — there are simply more, smaller ones. When analyzing any Korean dividend name, your first three checks should be: current record-date regime (legacy or reformed), payout frequency, and whether a formal dividend policy has been disclosed.

FAQ

When is the last day to buy a Korean stock and still receive the year-end dividend?

For legacy-calendar companies, two trading sessions before the register is fixed at year-end, because Korea settles T+2 and KRX does not trade on the final business day of the year. The exact date shifts with each year’s holiday calendar — confirm it on the KRX data portal rather than assuming.

Why does the stock price drop on the ex-dividend date if the dividend amount isn’t even announced yet?

The market prices in an expected dividend based on history, guidance, and analyst estimates. Because KRX does not adjust the reference price for cash dividends, the entire adjustment happens through trading. When the actual payout announced at the AGM differs from expectations, there is often a second, smaller repricing in February or March.

Why do Korean preferred shares trade at such large discounts?

Mainly the absence of voting rights in a market where control contests and governance events carry value, plus thinner liquidity. The discount has historically been wide — often 30–50% for major issuers — which is exactly why the yield pickup makes preferreds a recurring dividend-season rotation target.

Does foreign dividend-capture buying reliably lift the KOSPI in December?

No. It is one flow among several, it concentrates in index heavyweights rather than the broad market, and much of it is hedged with futures, so the net index impact is smaller than the cash buying suggests. Treat it as a flow to verify in the daily data each year, not a seasonal rule to trade blindly.

Sources

  • KRX data portal — record dates, ex-dividend data, investor flows: data.krx.co.kr
  • Naver Finance — daily foreign/institutional net-flow tables per stock and index: finance.naver.com
  • KOFIA statistics — customer deposits, margin lending: freesis.kofia.or.kr
  • Bank of Korea ECOS — FX and macro series: ecos.bok.or.kr

Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.


Related reading