KOSPI: 206% Up, 44% Down, and the Quarter Isn't OverKOSPI Composite IndexKRX:KOSPIIvanLabrieA country index has given back 43.93% from its high. The quarterly candle is down 34.01% and it does not close until the end of September. The structure behind that is the part worth your time. KOSPI spent twelve quarters building a base after the 2020 reset. The trend fired in Q1 2025, got knocked off course by the April tariff round, then fired again the following quarter. From there it ran three target ranges. The full advance measured 205.55%, roughly 3,072 to roughly 9,385. Three target ranges on a quarterly signal is already unusual. On a country index it is close to unheard of. The last time this chart did anything comparable was 2003. That signal ran into 2007 and delivered 171.98%, about 698 to about 1,897, and it topped in Q4 2007 on a target hit. Five quarters and 57% later it was over. This run was the bigger one. 205.55% against 171.98%, roughly 34 points more advance. The giveback has also been faster, 43.93% peak to trough inside a single quarter against five quarters last time. Why it broke this way is not a mystery. Two names carry this index, Samsung Electronics and SK Hynix, and SK Hynix is the cleanest HBM exposure available anywhere on earth. Leverage sat on top of that concentration. When the semiconductor repricing arrived there was nothing underneath to absorb it, and the same money that made the ascent steep had to sell. The leverage did not disappear when the index broke, it moved. SOXL took in about $2.4 billion since July began. In the week of July 3rd to 9th, five of the ten largest Korean net purchases of US securities were leveraged products. Some of the rest reached 50x Korean stock perpetuals on offshore venues. Capped leverage finds leverage somewhere else, on the same underlying, at a higher multiple, outside whoever is doing the capping. The authorities have now answered. The finance ministry is moving to stabilise the market and to cap retail exposure to leveraged ETFs. Nobody reaches for that particular instrument unless levered retail is the channel the damage travelled through. For the daily Korean tape underneath all of this, LoRosha's Investment Desk on Substack has been a better feed than anything on the wires. Today made the point better than I can. Samsung printed record quarterly earnings and guided to a memory shortage running to 2028, the stock was up more than 8% during the call, and the index still closed down 1.23%, because individual investors sold more than foreign and institutional money bought. That is what forced selling looks like when the fundamentals are not the problem. Two things are worth carrying out of this if you never trade Korean equities. The first is that the same shape exists elsewhere. An index carried by a handful of names, with retail leverage extended into them, is not a Korean peculiarity. Korea just got there first because it is the most concentrated and the most levered version of it. The US gave a version of the same tell this week. The Nasdaq index triggered its weekly downtrend and the S&P did not, and the difference between the two came down to one megacap's earnings reaction holding the broader index up. When a single company's report decides whether an index-level signal fires, the index has stopped describing the market and started describing that company. Same argument as Korea, on a market almost nobody files under concentrated. The second is about the semiconductor trade specifically. Levered Korean money did not rotate away from what hurt it. It went back into the same sector at three times the exposure. That is what the late stage of a crowded trade looks like from the inside. The numbers are all on the chart. The high this decline measures from sits near 9,385. This quarter opened at 8,591.50, printed 8,620.15, and has traded as low as 5,262.77. Now let me be exact about what this post is, because it would be easy to read it as a short call and it is not one. The quarterly signal here was a LONG. It fired off that twelve quarter base, it ran three target ranges, and those targets are where you take partial exits. There is no quarterly short signal on this chart, so I have no invalidation level to hand you, because there is nothing to invalidate. What I am doing here is observation. A country index that ran three ranges and then handed back 43.93% inside one quarter is abnormal enough to be worth writing down on its own. The trade signals on the way down lived on the daily, and they have already played out. So the open question is not a level I can give you. It is whether this rhymes with 2003 into 2007, where the advance ended on a target hit and the five quarters after it took 57%. Nobody can answer that from a quarterly bar with two months still to run, and I am not going to pretend otherwise. One last thing, because a post like this one invites exactly the wrong conclusion. A top being in does not mean the path from here is a straight line down. This is the same shape the US indices made at the 2022 top, and that top still produced a relief rally that ran six weeks, with eleven weeks passing from the start of that rally before new lows printed. That is a long time to be right about the direction and still be in the wrong position. What you do with that depends on your timeframe, your risk appetite and how you actually trade. If KOSPI is something you trade actively on the daily, this is where you watch for a daily entry to buy, or for continuation signals lower to keep shorting. Both of those are live from here, and I am not going to tell you which one to take. The quarterly chart tells you what kind of market you are standing in. Which of those two shows up next is a daily chart question. Macro and Time@Mode, same as always. Everything I publish is timestamped.