Monetisation is not announced, and no single series answers the question, since the informativeness of every candidate series depends on the state of the system being measured. The evidence tree exists because of that dependence, and the widget’s own headline scale is the first thing it demotes.
In a system with anchored expectations, growth of the money supply is itself the signal, since nothing else is generating it. In a system already monetising openly, the direct institutional markers stop discriminating — all of them are met — and the aggregates return as a measure of speed. Between those two states lies a transitional regime in which the aggregates are contaminated and the direct markers are not yet sufficient for an unambiguous reading.
That transitional state is the least studied of the three, and waiting for it to resolve itself buys time for the side that created it.
| Reg. | Name | Trigger condition | What sets the zone |
|---|---|---|---|
| I | Anchored | Observed inflation below 14% and the key rate net of inflation ≥ 0 | The quantitative scale |
| II | Transitional — current | Observed inflation ≥ 14% or the key rate net of inflation below 0 | Observed action and revealed preference — classes A and B |
| III | Open monetisation | Direct purchase of government debt by the central bank | The rate of money growth: aggregates return as a measure of speed |
The 14% boundary is the level at which, in Russian experience, expectations ceased to be anchored — observed in spring 2022. Calibrated on an episode, not chosen for roundness, which is the same procedure as in note 01.
Current state: observed inflation 15.1%, key rate net of inflation −0.7 to −1.1 pp. Regime II on both conditions independently.
| Class | What it is | Why the weight changes |
|---|---|---|
| A · Direct institutional | Actions the state takes: refinancing against state collateral, compulsory absorption of debt | Observed action. Strongest in any regime: the action has been taken and can be undone only by another action |
| B · Price and behavioural | The payer’s revealed preference: the real rate for the depositor, disintermediation of deposits | The payer bears a cost and therefore cannot lie. In regime II it carries the same weight as A |
| C · Quantitative aggregates | Money growth, the spread to nominal output | Demoted to confirmatory. An aggregate that improves when things get worse cannot set a zone — §4 |
| D · Control | Sterilising operations, the share of government debt in the repo collateral pool | Discriminators, not markers: they answer what the observed thing is, not whether it is observed. No zone weight in any regime |
The reading rule for regime II is published before the reading: orange at two or more markers across classes A and B, at least one of them from A; red requires both class A markers met in full; at one of two in class B the zone would be yellow. Current state — A: 1 of 2 with one more prepared and not fired, B: 2 of 2 — gives orange.
Class D is the one most often collapsed into the rest. Sterilising operations and the composition of the collateral pool do not tell you whether monetisation is happening, they tell you what the thing you are already observing is = essential to interpretation, useless as a trigger. They stay in the tree with the zone weight set to nothing.
The output deflator contains an export price component weighing roughly a quarter. In 2026 rouble export prices were falling, so the deflator comes out below consumer inflation entirely legitimately.
Thus if oil falls, the deflator falls, nominal output falls, the spread widens and the widget reddens — with the money supply unchanged. If oil jumps, the reverse. The denominator moves with the same exogenous variable the panel treats elsewhere as market conditions.
The scenario the panel expects would confirm the panel through arithmetic rather than through monetisation = the spread is demoted to a confirmatory marker.
This is the disclosure that costs something. The headline scale is the one quantity a casual reader takes away, and it has been stripped of authority by its own author.
The same discipline applies to the alternative denominator. The nominal output figure embedded in the budget law is published on the panel as a reference point and is not used, since it is the plan of the party whose monetisation the widget measures. A reconstructed denominator with a stated range is used instead, shown as a band rather than a line.
A reader who disagrees with the distribution can set their own weights and rerun the reading. Claiming a settled methodology here would be false precision.
Refutation of the reading, as distinct from the weighting: a reduction in banks’ liabilities to the central bank together with a falling floater share in new placements; the real rate for the depositor moving into sustained positive territory; an end to the outflow into cash alongside a rising term-deposit share.
Capital flight, sanctions evasion, credit stress and hidden defaults all have this shape.
Three failure modes.
An evidence scheme that cannot say which regime it is in will keep reading the last regime it understood.