On July 29 LG Electronics Inc. (KRX:066570) filed the result report for a treasury share buyback conducted through a specified money trust with NH Investment & Securities. The contract was for ₩100,000,000,000. The company spent ₩100,000,023,800 and acquired 730,429 shares.
The filing states there was no shortfall against the planned amount. This is a completed programme, not an abandoned one, and the termination is administrative.
What makes it worth reading is the month-by-month table, which Korean trust reports require and which most buyback disclosures do not give you.
Common share purchases, by month end.
February: 126,616 shares for ₩14,886,163,500. March: 139,081 for ₩16,000,359,000. April: 131,543 for ₩16,163,126,000. May: 109,167 for ₩21,535,853,300. June: 82,182 for ₩21,846,111,000. July: none.
The money going out rose. ₩14.89bn, ₩16.00bn, ₩16.16bn, ₩21.54bn, ₩21.85bn. Up 47% from the first month to the last.
The shares coming back fell. 126,616, then 139,081, then 131,543, then 109,167, then 82,182. June bought 35% fewer shares than February.
Divide one by the other and the implied average purchase price runs ₩117,570, ₩115,043, ₩122,874, ₩197,275, ₩265,825. The stock more than doubled while the company was buying it.
This is the third Korean buyback I have looked at this year with the same structure and the same outcome.
Hyundai Mobis budgeted ₩500bn in April against a share count calculated from the prior day's close, and by the time buying finished in July it had acquired 19.3% fewer shares than the plan implied, at an average 23.9% above the reference price.
The mechanism is identical. A company commits a sum of money over a window and lets the market determine how many shares it retires. In a rising market that guarantees the programme buys the most stock when it is cheapest and the least when it is dearest, which is the correct behaviour, and it also guarantees the announced share count is fiction.
Nobody is misled, because the disclosures are explicit about it. But an investor reading "₩100bn buyback" should understand that the share count is an output rather than a target.
Where LG Electronics differs from Mobis is that it started early and the whole programme is comfortably in the money.
Blended across the five months, the company paid ₩153,642 per common share. The August 27 close was ₩198,300. So the shares bought are worth 29.1% more than what was paid for them.
February and March, at ₩117,570 and ₩115,043, were bought at roughly 59% of the current price. June's purchases at ₩265,825 are 25.4% underwater. On balance the programme created value, and that is largely because it began before the share price re-rated rather than after.
That timing is not luck exactly, but it is not skill either. A company that decides in January to spend ₩100bn over six months has no idea what its shares will do in May. The lesson is about the design, not the outcome.
One detail worth pulling out. Of the ₩100bn, ₩90,431,612,800 went into common shares and ₩9,568,411,000 into preferred.
The preferred purchase bought 141,840 shares, an average of ₩67,459 each. That is 44% of the ₩153,642 average paid for common.
LG Electronics' preferred shares carry the same economic claim on earnings as the common but no vote, and in Korea they routinely trade at large discounts for that reason. A won spent on preferred therefore retires more than twice as much earnings entitlement as a won spent on common.
The company allocated roughly 9.6% of the programme to preferred, against 17,181,299 preferred shares outstanding versus 162,884,638 common, so about 10.5% of the count. That is a proportional split rather than a value-maximising one.
Hyundai Motor took the other approach in August, cancelling 2.003% of its preferred class against 0.631% of its common, weighting three to one toward the cheaper shares. On pure arithmetic that is the better use of the money, and it is a choice available to LG Electronics too.
The last and most important distinction. These shares are in treasury. The filing records holdings after termination of 588,589 common, or 0.36% of the class, and 141,840 preferred, or 0.83%.
Treasury shares are not the same as retired shares. They cannot be voted and they do not receive dividends, so the economic effect on continuing shareholders is similar while they sit there. But they can be sold again, used for employee compensation, or placed with a third party, at which point the benefit reverses.
NAVER cancelled 4,901,094 shares outright in July, taking them permanently out of existence. Hyundai Motor did the same with 2,505,606 in August. LG Electronics has bought and held.
There may be good reasons. A company with employee stock plans needs treasury shares to satisfy them. But an investor counting the ₩100bn as a return of capital should note that it has not yet been made irreversible.
A cancellation resolution. Korean boards can retire treasury stock at any time out of distributable profit, and doing so takes about a day. If one appears in the coming months, the ₩100bn becomes a permanent reduction in the share count. If the shares sit in treasury through 2027, the programme was a market intervention rather than a capital return.
The second thing is whether another trust contract follows, and how it is denominated. Every Korean buyback I have examined this year was sized in won over a fixed window. A programme sized in shares, or one with a price condition attached, would be the first sign that a Korean board has noticed what the fixed-sum structure costs when the stock is rising.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.