Skip to main content
Seasonality Investing

Seasonality Investing · Stock Seasonality, 8 Lessons · Lesson 6

Korea's market calendar was made by taxes

KOSDAQ's average return was negative in all five months from June to October, while January's average was +4.62%. See in numbers the tax rules that made this calendar, and how it changed as the rules changed.

15 min readLesson 6 / 8As of: KOSPI Feb 1997–Sep 2026 · KOSDAQ Nov 2000–Sep 2026 · calculated 2026-10-10✓ Read

QToday's question

Why does news about a flood of individual-investor selling appear every year-end?

Every December, the phrase 'selling to avoid the major-shareholder capital gains tax' comes up. This lesson starts with one KOSDAQ number, looks at the tax rules and data behind it, and ends with the twist from changing rules and one action to try today.

MAPLearning map for this lesson

We build it up in three steps. At the end of each step, a 'Quick check' lets you test your understanding before moving on.

Before you start

In Lesson 1, Why is September so hated?, we learned how to read the four numbers of mean, median, win rate and sample size, and in Lesson 3, Santa comes, and January speaks, we looked at the US year-end effect. If small caps were the star of the US January effect, in Korea the tax rules are the star. The KOSPI half-year numbers from Lesson 2, Sell in May, come up again here.

  1. 1 · One numberFive months in a row in the red

    You can remember KOSDAQ's June–October and January numbers together with their periods and sample sizes, and explain why January's mean and differ so much.

  2. 2 · Insight · 3 · DataA calendar made by taxes

    You can explain how the fact that the major-shareholder capital gains tax is judged by holdings at the end of the fiscal year creates a flow of December selling and January rebuying, and read the KOSPI and KOSDAQ monthly tables.

  3. 4 · Twist · 5 · ActionRules make it and rules change it

    You can separate the fact that KOSDAQ's January excess return disappeared after 2016 from its candidate causes, and check December individual net selling yourself on KRX.

Building knowledgeConcepts built so far: 29 · 5 added in this lesson

Seasonality = tilt in probabilityMean, median, win rate, sample sizeComparing against a baselineSince 1950 vs the last 20 yearsNot 'sell' but 'lower your expectations'Nov–Apr vs May–Oct half-year comparisonHalloween effect (36 of 37 countries)The gap vanishing over the last 10 yearsFixed dates vs MACD-confirmed switchingSanta rally's 7 trading days vs the baselineJanuary barometer 2×2Small-cap January effect (excess return)The more famous the pattern, the sooner it weakensYear-end/new-year = a filterTurn-of-month effect 90% → 23%When the structure changes, the pattern disappearsThe shorter the cycle, the sooner it weakens4-year presidential cycle and a sample of 18–19One-line cause testHalf-year sector rotation (Stovall)Seasonality created by real demandExcess return vs the market and number of winsTech exception, defensive reversalMonthly gold futures record and the limits of continuous futuresKOSDAQ weakness in Jun–Oct, strength in JanuaryNEWMajor-shareholder capital gains tax and the year-end record dateNEWFading of the January excess return (since 2016)NEWPatterns made by rules are changed by rulesNEWChecking December individual net selling yourselfNEW

1 · One numberFive months in a row in the red

June to October, all five months −Months when KOSDAQ's average monthly return was negative2001 to Sep 2026, 25–26 observations per month · price index basis

Lay out KOSDAQ's average monthly returns since 2001 and all five months of June (−2.25%), July (−1.00%), August (−0.66%), September (−1.88%) and October (−0.98%) were negative. Conversely, January's average of +4.62% was the highest of the twelve months. Going by the numbers alone, the claim "KOSDAQ is weak in summer and strong in January" seems plausible.

But we have to read it again with the four numbers from Lesson 1. January's is +1.79%, less than half the mean. That is because a single January, in 2001, returned +60.4%. Without that one year, January's average falls to +2.39% (25 observations). And with only 25–26 observations per month, it cannot be read with the same weight as the US S&P 500 (76–77 per month).

Quick checkBefore taking "KOSDAQ January average +4.62%" at face value, what else should you look at?
The median (+1.79%), the sample size (26) and the . A single +60.4% in January 2001 pulled the average up sharply. Without that year, the average is +2.39%.

2 · InsightA calendar made by taxes

Korea's year-end and new-year pattern has a clearer candidate cause than the US one: the . Capital gains on domestically listed stocks are not taxed unless you are a major shareholder, but a major shareholder who holds a certain amount or more of a single stock pays capital gains tax when selling. And whether you are a major shareholder is judged by your holdings at the .

So individual investors near the threshold have an incentive to sell part of their holdings before the end of December to get below it. It is said that many buy again once the new year starts. Etoday (2019-12-02) reported that individual net selling in December was KRW 5.13 trillion in 2017 and KRW 1.58 trillion in 2018. Many explanations say this flow is larger on KOSDAQ, where individual investors have a bigger share.

Easy analogy · The scale on December 31

If your weight at a health checkup were recorded only on December 31, some people would starve briefly before that day and eat again in January. The explanation is the same here: the major-shareholder test is set by holdings on a single year-end day, so trades that 'cut back briefly and then rebuild' cluster just before it.

How the major-shareholder threshold has changed

Click a year. The threshold at that time and the reason for the change will appear.

Holding-value threshold per stock (KOSPI basis; KOSDAQ shown separately). Compiled from Korea Economic Daily (2025-09-15), Newspim (2025-08-06) and the Democratic Party briefing (2023-12-21). Ownership-percentage criteria and detailed requirements are omitted, and you should check National Tax Service guidance for actual application.

This pattern is a classic case of the seasonality with a cause mentioned in Lesson 1. Because the cause is built into the rules, it can be explained, and in return you can expect the strength of the pattern to change when the rules change. However, concentrated selling does not necessarily push the index down. That is because who absorbs the selling also plays a role.

Quick checkIf the major-shareholder test changed from year-end holdings to the average holdings over the year, what would happen to the year-end selling pattern?
There is less reason to sell all at once at year-end. Since dipping below the threshold for just one day would no longer avoid the tax, you can infer that the December selling and January rebuying flow would weaken (this is an inference based on the assumption).

3 · DataKOSPI and KOSDAQ 12-month report card

We calculated the monthly records of KOSPI (from Feb 1997) and KOSDAQ (from Nov 2000) in the same way. The baselines are an all-months average of +0.97% and a of 55.1% (356 months) for KOSPI, and an average of +0.38% and a win rate of 53.1% (311 months) for KOSDAQ. Also keep in mind that the win-rate baseline is lower than in the US (average +0.75%, win rate 60.4%).

KOSPI and KOSDAQ monthly report card

KOSPI (29–30 observations per month)KOSDAQ (25–26 observations per month)
Tap a bar.

Calculated directly from month-end values of Yahoo Finance ^KS11 and ^KQ11 daily closes (price index, excluding dividends and costs). KOSPI Feb 1997–Sep 2026, KOSDAQ Nov 2000–Sep 2026. Retrieved 2026-10-10. The dotted line in the win-rate view is each index's all-months win rate.

KOSPI had high averages in January (+3.21%, 29 observations), April (+3.18%, 30) and November (+2.91%, 29), and negative averages in August (−0.83%), September (−1.13%) and October (−0.43%). KOSDAQ was negative in all of June–October. However, KOSDAQ's September has an average of −1.88% but a median of +0.11%, which shows that a few large declines dragged the average down.

Grouped by half-year, the difference gets bigger. As we saw in Lesson 2, over 29 years in 1997–2025 KOSPI averaged +12.54% in Nov–Apr (win rate 69%) and −0.15% in May–Oct (win rate 41%). KOSDAQ over 25 years in 2001–2025 returned +10.64% (68%) in Nov–Apr and −4.94% (36%) in May–Oct. But these are observations taken once a year, fewer than 30 in all.

Quick checkKOSDAQ's September has an average of −1.88% and a median of +0.11%. How should you read it?
It is closer to 'a few big declines dragged the average down' than 'most Septembers were very weak.' The median is near zero, so about half were positive. The actual win rate is also 50%.

4 · TwistThe January excess return disappeared after 2016

The twist. KOSDAQ's January (vs KOSPI) averaged +2.67 percentage points over 2001–2026, but looking only at the 11 years since 2016, it flipped to −0.47 percentage points. Yet in 2016–2020 the major-shareholder threshold was actually lowered from KRW 2.5 billion to KRW 1.5 billion to KRW 1 billion, so more people became taxable. Taxes alone do not explain it.

KOSDAQ − KOSPI, excess return by year

Click a bar. That year's KOSDAQ and KOSPI returns will appear.

Excess return = that month's KOSDAQ return − KOSPI return (percentage points). Calculated directly from Yahoo Finance ^KQ11 and ^KS11 month-end closes. January 2001–2026 (26 observations), December 2000–2025 (26 observations). Gray bars are 2001–2015, blue bars are 2016 onward.

Looking at the bars, the big January excess returns are concentrated in a few years such as 2001 (+38.0 percentage points) and 2005 (+20.2 percentage points), and since 2016, apart from 2018 (+10.4 percentage points), positives and negatives have alternated. December flipped the other way, from an average of −1.62 percentage points in 2001–2015 (15 observations) to +0.71 percentage points since 2016 (10 observations). That means KOSDAQ has been less weak at year-end and less strong in January.

There are several candidate causes. Some point out that in December, when individual selling is concentrated, institutions and foreign investors buy that supply cheaply and sell it back in January (Hankyoreh column, 2025-08-31). It is also possible that other investors trading a widely known pattern earlier pulled the effect forward and weakened it. The December 2023 threshold being raised again to KRW 5 billion, which reduced the number of sellers, is another recent factor. Since none of these can be confirmed, we treat all of them as hypotheses.

Common misconceptions

"Individuals sell at year-end, so KOSDAQ must fall in December?" No. If other investors absorb the selling, the index holds up. Since 2016, the average December excess return has actually been positive.

"Buy KOSDAQ in January and you beat KOSPI?" Since 2016, it won only 5 times out of 11. Split the period and the numbers look completely different.

"The threshold stayed at KRW 5 billion, so the pattern comes back?" There are only a few years of sample to judge by. A rule change is something to check, not a signal.

Quick checkThe major-shareholder threshold was lowered in 2016–2020, yet the January excess return disappeared. What does that tell us?
That there may be more than one cause. Even if taxes create selling, the index result changes depending on who takes the other side and how widely known the pattern is. That is why even a pattern with a cause should be rechecked over a recent period.

5 · One actionCheck December individual net selling yourself

This lesson's action is simple. On the Korea Exchange (KRX) information data system, find December individual net selling for the last 5 years yourself, write it in the table below, and place it next to that year's KOSDAQ December and January results. Five numbers you checked yourself will stay with you much longer than someone else's 'year-end selling bomb' article.

Practice · Find December individual net selling on KRX

  1. Go to the KRX information data system (data.krx.co.kr) and open 'Statistics → Basic Statistics → Stocks → Trading Performance → Trading by Investor Type.'
  2. Set the market to 'KOSDAQ' (or KOSPI) and the query period to December 1 through the last trading day of December of that year.
  3. Set the unit to 'trading value' and check the number in the individual 'net buying' cell. A negative number means net selling.
  4. Repeat for the last 5 years in the same way and enter the figures in the boxes below, in units of KRW 100 million.
Screen example · Trading by investor type
Market: KOSDAQPeriod: 12/01 ~ end of 12Unit: trading value
Institutions totalOther corporationsIndividuals · net buyingForeign investors total
Enter the numbers and bars appear, along with that year's KOSDAQ December and January returns.

Menu names and locations may change as KRX redesigns its screens, and the picture above is a screen example, not the actual screen. The example values are the combined KOSPI and KOSDAQ individual net selling reported by Etoday (2019-12-02), which may differ from the market selected in the input boxes. What you enter is stored only in this device's browser.

After filling in the numbers, check two things. First, whether KOSDAQ's December result was actually bad in the years with large net selling. Second, whether there was a rebound in January of the following year. Five years cannot settle anything, but it clearly shows that "the rules create selling" and "selling pushes the index down" are two different statements. You can also see investor-type flows for the whole market on the flows page.

The figures in this lesson are based on price indexes, so dividends, trading costs and taxes are left out. The major-shareholder threshold and the way it is judged can change again with tax law revisions, so check National Tax Service guidance before making any real decision. For educational purposes only, not investment advice.

Quick checkIn one year, December individual net selling was very large, yet KOSDAQ's December return was positive. What should you check?
Who took the other side. If you look at net buying by institutions and foreign investors on the same screen, you can tell whether other investors absorbed the individuals' selling.
The one sentence left from Lesson 6Rules make it and rules change it

✓What we learned today

  • Since 2001, KOSDAQ's average was negative in all five months from June to October (25–26 observations per month). January's average was +4.62%, but the median was +1.79%, and a single +60.4% in January 2001 pulled the average up.
  • The major-shareholder capital gains tax decides who is subject to it by holdings at the end of the fiscal year (end of December). That is the explanation given for the flow of individual selling in December and rebuying in January. December 2017 individual net selling was KRW 5.13 trillion.
  • KOSDAQ's January excess return (vs KOSPI) averaged +2.67 percentage points over 2001–2026 (26 observations), about +5.0 percentage points in 2001–2015, and −0.47 percentage points since 2016 (11 observations).
  • The twist: in 2016–2020 the major-shareholder threshold was actually lowered from KRW 2.5 billion to KRW 1 billion, so more people were taxable, yet the January excess return disappeared. Taxes alone do not explain it, and the possibility (a hypothesis) that other investors traded the known pattern earlier is also raised.
  • In September 2025 it was decided to keep the major-shareholder threshold at KRW 5 billion. A pattern made by rules changes when the rules change. So always check rule changes and sample sizes together.
Continues in the next lesson · Lesson 7Why do patterns die? So how do you use them?

Today we looked at Korea's calendar, where the cause, taxes, is clear. In Lesson 7 we cover why patterns die, and how to use seasonality as a filter with the three-line 'condition · action · invalidation' rule.

Quiz5-question quiz

Pick an answer and it's graded right away with an explanation.

Q1. Holdings at which point mainly decide whether you are subject to the major-shareholder capital gains tax?

Q2. KOSDAQ's January average is +4.62% and its median is +1.79% (2001–2026, 26 observations). What is the main reason for the large gap?

Q3. How did KOSDAQ's January excess return (vs KOSPI) change after 2016?

Q4. The major-shareholder threshold was lowered from KRW 2.5 billion to KRW 1 billion in 2016–2020, yet the January excess return disappeared. What is the lesson?

Q5. What is the most suitable action for this lesson's assignment?

NextTerms · Next lesson · Related tools

Major-shareholder capital gains taxLarge-shareholder capital gains tax

Unless you are a large shareholder, there is no capital gains tax on profits from trading domestically listed stocks, but a large shareholder holding more than a set amount in a single stock pays capital gains tax when selling. The per-stock holding threshold was lowered from KRW 5 billion in 2013 to KRW 1 billion in 2020, raised back to KRW 5 billion in December 2023, and in September 2025 it was decided to keep it at KRW 5 billion.

Year-end holdings testYear-end holding test

A method that decides large-shareholder status by the amount held at the end of the prior fiscal year (usually the end of December). Because that single day's holding is the benchmark, the explanation goes that selling piles up just before year-end as holders sell part of their stake to drop below the threshold.

Tax-loss sellingTax-loss selling

Selling stocks that have lost money near year-end to lock in the loss and reduce that year's taxable gains. It is often cited as a cause of year-end weakness and the rebound at the start of the year.

MedianMedian

The value that falls in the middle when values are lined up by size. It is less affected by one or two outliers, so it is used to check whether the average has been inflated.

OutlierOutlier

A value that is far larger or smaller than the others. The average is pulled hard by a single outlier, so when the average and median differ widely, first check whether an outlier is mixed in. Example: KOSDAQ, January 2001, +60.4%.

Sample sizeSample size (n)

The number of observations used to calculate a statistic. In monthly seasonality it means how many times that month has occurred; the smaller the number, the more likely the result is due to chance.

Excess returnExcess return

An asset's return minus the return of a benchmark. For example, if the Russell 2000 returned +3% in January and the S&P 500 returned +2%, the January excess return is +1 percentage point.

Put what you learned to use right away

Foreign & Institutional Flows→Overseas stock capital gains tax calculator→

For educational purposes only, not investment advice. Past statistics do not guarantee future returns. KOSPI and KOSDAQ figures were calculated directly from month-end values of Yahoo Finance ^KS11 and ^KQ11 daily closes (price index, excluding dividends and trading costs; retrieved 2026-10-10). The history of the major-shareholder threshold is cited from Korea Economic Daily (2025-09-15), Newspim (2025-08-06) and the Democratic Party briefing (2023-12-21), and December individual net selling from Etoday (2019-12-02).