Why did the pension fund that always bought Korea's trading halts sell into five of them this year?

When Korean stocks fell hard enough to halt trading, the National Pension Service used to be the buyer. Six circuit breakers were triggered between the system's launch in December 1998 and the end of 2025, and the fund put money to work in all six. 2026 broke the pattern. Of this year's halts, the fund chose to sell rather than buy on five of them.

Bourse operator issues circuit breaker for KOSPI on sharp fall (The Korea Times, 2026-07-29)

출처: koreatimes.co.kr

The exchange halted trading for 20 minutes at 12:31 p.m. on July 29 after the KOSPI fell more than 8% from the previous close. The exchange counted that halt as the eighth of the year. The original post counts nine, and the gap appears to come from whether KOSDAQ halts are counted alongside.

Why the fund could not buy

The National Pension Service fixes in advance how much of the total fund domestic equities may occupy, and when a rally pushes the holding above that share, it sells to bring the weight back down. That is rebalancing. The part that matters is not the selling but what the selling produces. Cash trimmed during a rally is the cash available to buy during a crash. Think of a shopping basket with a set ratio per compartment: whatever you take out of an overflowing compartment is what you spend next time.

The rule changed in May

On May 28 the fund's management committee raised the domestic equity target from 14.9% to 20.8%. The change answered a complaint that mechanical selling into every rally was pressing down on the market.

NPS raises domestic stock target to 20.8% amid market pressure (Seoul Economic Daily, 2026-05-28)

출처: en.sedaily.com

Seoul Economic Daily reported that the actual domestic equity weight sat near 30% when the decision was made. Lifting the target by 5.9 percentage points still left the real holding more than 9 points above it. Even with a wider allowance, the fund stayed on the selling side.

The same fact splits on whether you read it as a management miss or as the back side of a rule changed to help the market.

NPS sits out Korea rout as wider equity bands blunt rebalancing (KED Global, 2026-07-29)

출처: kedglobal.com

KED Global argues that the wider allocation band, introduced to curb forced selling into rallies, also dulled the fund's ability to buy aggressively into a collapse. The same article notes that the fund made only modest purchases even on July 28, when the KOSPI closed 10.84% lower at 6,023.66.

What the missing 219 trillion won was

At a KOSPI of 9,100 on June 19, the fund's year-to-date paper gain stood at 314 trillion won. With the index closing at 5,663 on July 29, the same calculation gives 95 trillion won. A paper gain is the amount written into the books for an asset that has risen in price but has not been sold. It is not money in hand. It is an estimate of what the money would be if sold. So when the price retreats, the entry is erased with it. By the original post's arithmetic, delaying the fund's depletion date by one year costs 49 trillion won. That makes 314 trillion won worth 6.4 years and 95 trillion won worth 1.9 years.

How we will know

In the end there is one question. Is the buying that resumed in late July the safety net returning, or simply what falls out once the weight has drifted back under the cap. The 116 billion won bought on July 28 and 338 billion won on July 29 are far larger than the 48 billion won of June 8 or the 29 billion won of July 7. Set against the 573 billion won of March 13, 2020, they are still small. If the next halt brings buying in the trillions of won, the constraint has genuinely loosened. If the figure falls back to the low hundreds of billions, this week's purchases were the mechanical result of a weight that had already come down. Acting as the market's safety net was never the fund's objective. The objective is return, and the safety net was a by-product of chasing it. What remains is that this year the by-product went missing.