The threshold, and the slope underneath it
Household credit in South Korea stood at 1,993 trillion won at the end of March, up 14 trillion won from the end of last year and a seventh consecutive quarterly increase. Outstanding household loans across all financial institutions are estimated to have risen by at least 21 trillion won in the second quarter, which carries the total past 2,000 trillion won, or 1.41 trillion dollars, for the first time. The arithmetic itself is undramatic. What deserves attention is that the pace of growth slowed for a second straight quarter while the level kept climbing: a system taking on debt more slowly, and still taking it on.[1]
The composition is where the question sits. Government measures aimed at cooling the housing market have not weakened demand for homes, and loans taken out to buy shares have climbed alongside the stock market rally. Across 2025 as a whole household credit grew 56.1 trillion won, or 2.9 per cent, the fastest since 2021. Two kinds of collateral are therefore doing the work at the margin, and they behave differently when the weather turns. A home's price is settled when it changes hands. A share is priced at every session.[1]
What that collateral did last month
The second kind of collateral has just had a violent quarter. The Korean market crashed last month on doubts about the profitability of AI-related investment and what those doubts mean for chip demand, then rebounded: the Kospi is up 6 per cent this month and the tech-heavy Kosdaq 20 per cent. Short selling did not retreat with the recovery. Outstanding short interest reached about 19 trillion won, or 13.4 billion dollars, on Tuesday, 14 per cent above the 16.73 trillion won of a month earlier, on Korea Exchange data. The series has swung from 15 trillion won at the end of February to 12 trillion in early March and 23 trillion in early June.[2]
Put the two together and a transmission channel appears that a mortgage book does not have. When borrowing is secured on shares, a drawdown reaches the household balance sheet at the speed of the closing price: collateral values fall, top-ups are demanded, and the cheapest way to meet them is to sell the very shares that fell. July showed the market can travel that far in weeks, and the rising short interest says a meaningful group of participants does not treat the rebound as settled. The other reading deserves saying out loud: short interest rises for hedging and arbitrage as much as for direction, and the data do not disclose how much of the 21 trillion won is share-backed rather than home-backed. If housing still dominates the increase, the equity-sensitive slice may be small enough to absorb.[1], [2]
Buffers, and the test
Several buffers argue against alarm. Growth in household credit has slowed for two consecutive quarters, so the flow is decelerating even as the stock reaches a new high. The cooling measures are already in force. Short interest at 19 trillion won sits below the 23 trillion won of early June, so positioning is not at its own extreme. And last month's selloff came and went without reported strain in household credit, which is evidence that the buffer held once at that amplitude. None of that bounds the exposure. It bounds how loudly anyone is entitled to talk about it.[1], [2]
So here is the falsifiable version. If the Kospi gives back this month's 6 per cent advance before the third quarter closes, the quarterly increase in household credit should come in below the 21 trillion won estimated for the second quarter, and the slowdown should concentrate in share-backed lending rather than in housing loans. If the total instead keeps rising at the second quarter's pace through an equity drawdown, the share-backed portion is smaller than this reading assumes and the collateral channel is the wrong thing to watch. Either way, the third-quarter household credit figures answer it.[1], [2]