Regulation is not weather. It is a sequence of dated decisions.

The carbon market talks about regulation the way it talks about weather: a general condition, gradually worsening, that everyone should probably prepare for. This is precisely the wrong mental model, and it is costing holders money. Regulation does not arrive as a climate. It arrives as decisions, each taken on a specific day by a named body and published in a document you can download. Each one reprices standing positions immediately, silently, and without any of the machinery that would tell a holder it had happened.

The distinction is not semantic. A trend cannot be recorded, dated or priced. An event can.

Three ways a position loses value without anything happening to it

Start with what makes this class strange. Most assets lose value because something happens to the asset. Carbon has three loss mechanisms that leave the underlying tonne entirely untouched.

Eligibility withdrawal. A credit that could be surrendered into a compliance regime, and now cannot, is a different asset. No forest burned. No fraud was found. A committee published a list and a use disappeared. The largest example is running right now: of roughly 980 million legacy credits requesting transition into the Paris Agreement's crediting mechanism, around 13 per cent have host-country approval, with the two largest historical supplier countries absent from the approval list.

Standard revision. A corporate net-zero standard that stops recognising a category for target accounting does not invalidate the credits. It removes the reason the buyer purchased them. The book bought to hit a target becomes a book with no target to hit.

Methodology reassessment. A methodology withdrawn or downgraded reprices every vintage issued under it, including credits already sitting on balance sheets, in a single decision.

The repricing has already happened, and it is enormous

If regulation were a gradual condition, prices would drift. They have not drifted. They have bifurcated.

Note that this chart requires a logarithmic axis to be legible at all. That is the finding. A market where the strongest and weakest instruments differ by two orders of magnitude, while sharing an accounting unit, is a market in the middle of a reclassification.

The correct description of what is happening to avoidance credits is narrower and more useful than the one usually offered. Avoidance is not dead. A compliance scheme's second phase re-approved jurisdictional avoided deforestation, and integrity-labelled REDD has a pipeline in the hundreds of millions of tonnes. What has happened is a demotion: from neutralisation asset, usable against a target, to transitional compliance asset, usable in narrower and specified circumstances.

Demotions can be dated. Deaths cannot be traded around. The difference matters to anyone holding the stock.

Why nobody tells you when it happens

Here is the structural point, and it is the one worth taking away.

In credit markets, a downgrade arrives with machinery attached. There is a watchlist before it. There is a published methodology explaining it. There is a notification when it happens. There are decades of migration statistics telling you how often a bond of this rating moves down a notch, so the event can be provisioned for before it occurs.

Carbon has none of that. There is no watchlist, no negative outlook, no notification and no migration statistics. A methodology is eligible or it is not, a category is in a list or it is out, and when the line moves it moves for everyone holding on the wrong side of it, simultaneously and without warning.

A holder discovers a repricing the same way the public does. For a position that will be measured at fair value every reporting date from FY2028, that is an uncomfortable operating model, and it is the one currently in use across the market.

The denominator problem, which makes it worse

Even where the market attempts to quantify one of these events, it cannot agree with itself. The same eligibility loss has been published three ways.

Two hundred and thirty-six million tonnes and 32 per cent. Around 240 million and roughly a third. Two hundred and sixty-eight million and 30.6 per cent. Every one of those is defensible, and the difference between them is entirely in the denominator, which none of them stated.

That is a thirty-two-million-tonne spread on a single event, available to anyone who wants to select the framing that helps. In a market where the same event can be quoted three ways, a percentage without its denominator is not information. It is a position.

What a holder should actually do

Four things, none of which requires a view on where regulation is heading.

Treat eligibility as an exposure, not a status. Every position should carry the scheme, the methodology and the vintage that determine its eligibility, and those attributes should be monitored rather than recorded once at purchase.

Know your dates before they arrive. The decisions on the chart above are published. A holder who does not know which of their positions each one touches will find out when everyone else does.

Demand denominators. Any figure quantifying an eligibility loss on your book should state its numerator, its denominator and its as-at date. If it cannot, it is a rhetorical device rather than a measurement, and your auditor will eventually establish this on your behalf.

Ask when the method gets set, not whether. Fair value measurement of these positions binds on a published timetable. The technique will be chosen once and then lived with. It should be chosen deliberately and in advance rather than assembled under audit pressure, and it should not be chosen from a broker's quote.

The record is the answer, and it is a boring one

There is no clever instrument that solves this. What is missing is the least glamorous artefact in financial infrastructure: a list.

Every eligibility decision, dated. Every standard revision, with the categories it moved and the population it touched. Every methodology reassessment, joined to the projects and vintages actually affected. Each entry resolving to the document that decided it, each ratio carrying its denominator, and the whole thing free to read, because a record people have to pay for cannot become the reference the market prices against.

That is what a watchlist is, before anyone gives it a grander name. Credit markets built one, then built migration statistics on top of it, then built the instruments that migration statistics make possible. The order was not accidental.

Carbon is at the beginning of that sequence, holding several hundred billion dollars of ambition and no list.