The Label That Moved

One of five labels moved in this reading. Marker 5, gold's return to a settlement role, reads drifting as of July 30, down from on-track in May. The direction of central bank gold accumulation did not reverse; the second quarter of 2026 set a record for second-quarter purchases, bought into a 16% price decline. What changed was the evidence the May reading stood on. The World Gold Council cut its first-quarter purchase estimate from 244 tonnes to 57 after reclassifying flows to the over-the-counter market, which removed the headline number both the Marker 2 and Marker 5 readings had used. The IMF's July 1 reserves brief stated that gold's overtaking of US Treasuries in 2025, the crossover Appendix G named as a leading indicator, was driven almost entirely by price valuation. And the settlement function the marker is named for did not advance: the one prototype remains a research exercise with no reported volume. Accumulation held. The case built on it got weaker. The label says so.

Drifting is not a softer word for wrong. The methodology defines it as directional evidence that partially supports the marker, with caveats, and that is a precise description of this reading for Marker 5. The temptation in a project like this is to treat labels as reputational—to hold on-track as long as the headline direction allows, and to reserve movement for outright reversal. That would make the labels decorative. A reading is only as good as the evidence underneath it, and when a source revises a figure by 77% after publication, or an official institution attributes your leading indicator to price rather than to buying, the honest response is to say the ground moved, even if the structure on top of it is still standing. Drifting records exactly that. It also states what would move the marker back: a gold-anchored settlement instrument with reported volume, or a second consecutive record quarter of official-sector buying with the valuation caveat no longer doing the work.

The May essay said this could happen. It said several markers could shift to drifting, unclear, or reversed by July if conditions changed. One did, and two others carried corrections. The Marker 3 reading in May projected CIPS settlement would reach roughly $30 trillion in 2025; the actual figure was $25.55 trillion, up 2.7%, and the data was available when the May reading was written. That was an error of research, not of revision, and the July reading says so. The Marker 2 reading in May compared a single quarter of gold purchases to the entire sovereign Bitcoin position; the quarter it used was the one later cut by three-quarters. The comparison survives on the revised numbers—the second quarter alone still exceeds the US government's Bitcoin holdings—but it survives on different numbers than it was made with, and the reading names them. None of this is the framework failing. It is the only mechanism by which a framework published in public earns the right to be read a year later.

Four markers held, and one of them received the strongest kind of evidence a marker can get. On June 1 the Central Bank of Nigeria published a strategy document that describes the eNaira's adoption as slow, puts its cumulative transactions since 2021 at roughly ₦22 billion—about $16 million—and proposes sustaining it through government disbursement rather than consumer choice. The issuing institution documented the gap the marker predicted. The Atlantic Council's tracker, updated in May, stated the pattern in its own summary: advanced economies retreating from retail CBDCs, emerging markets pressing on in response to dollar stablecoins. Stablecoin supply saw its largest monthly contraction since 2022 in June while transaction volume set a record; supply fell, use did not. CIPS settlement held flat against the prior quarter while the renminbi's reserve share stayed under 2%. The decoupling claim is about the relationship between the series, and the relationship held, though the reading records that this was a consolidation quarter and that the dollar's reserve share rose.

I grew up in a country where the official currency and the currency people actually saved in told different stories for years before the sucre was abandoned. The book has a term for continuing to tell the old, legible story after the constraints behind it have eroded: narrative arbitrage—profiting from trust memory to cover a present the numbers no longer support. A site that tracks a framework can do the same thing to its readers. It can hold labels steady because movement looks like weakness, cite the original figure because the revision is inconvenient, and let the prior reading quietly disappear when the new one goes up. The July readings do none of those. Past readings are now preserved in full on each marker page, as published. Revisions are disclosed in the reading that follows them. Falsification conditions are carried forward verbatim, and where one registers evidence, the reading says which.

The next reading will test whether this was a pause or a turn. For Marker 5, two conditions are stated above. For Marker 3, the question is whether CIPS flatness and the dollar's reserve-share gain extend into a second quarter—which the reading's own conditions say would be a sustained registration, not a single data point. For Marker 4, whether the June supply contraction was a month or a trend. The framework's claim was never that these markers would read on-track forever. It was that their direction could be read, in public, against evidence that anyone can check, with the conditions for being wrong stated before the evidence arrives. One label moved. That is the system working. The next quarterly reading is October 30, 2026.