What a central bank balance sheet tells you that the rate does not
The judgement calls on top of the mechanics — and the three situations where the framework should be abandoned rather than used.
Key points
- Expansion and contraction are not symmetric signals: runoff shrinks a portfolio without anyone deciding to tighten.
- Two central banks with identical balance sheet sizes can have opposite policy content.
- Funding stress, fiscal dominance and small-open-economy effects each defeat the framework; the honest answer then is to say nothing.
This is a companion to our research note on reading central bank balance sheets. The note sets out the mechanics. This one is about the judgement calls that sit on top of them — the places where an analyst can talk themselves into a position the data does not support.
The asymmetry that flatters the signal
An expanding balance sheet is a commitment. A contracting one is often an accounting artefact of maturities. When reinvestment ends, the portfolio shrinks mechanically until someone actively sells, which almost nobody does. So “the balance sheet contracted” is weak evidence of tightening intent, while “it expanded” is strong evidence of easing intent.
Any framework that weights the two signals equally will over-read tightening and under-read easing. That is a modelling choice, and it is usually made silently.
Composition beats size, every time
Consider two central banks with portfolios of identical size. One holds short-dated bills. The other holds long-dated bonds acquired years ago and is simply letting them run off. Their published balance sheets have the same headline number and opposite policy content.
This is the single most common analytical error we see in commentary on quantitative policy: quoting a total and reading intent into it. The composition tells you what was decided; the total only tells you what has happened to happen.
When the framework is the wrong tool
Three situations defeat it, and in all three the honest response is to say so rather than to produce a number.
- Funding stress. Expansion as market plumbing. Intent is absent from the signal entirely.
- Fiscal dominance. When a central bank is absorbing a domestic debt issuance that keeps rates from rising, portfolio growth reflects the fiscal position, not an independent monetary choice.
- Small open economies. Foreign holdings of domestic currency contaminate the liability side and, through it, the reading of intent.
The practical discipline
State the episode count behind any timing estimate. Ours rests on a handful of post-2008 cases; that is a prior, not a finding, and it would be dishonest to quote it without that caveat attached.
Then name the observation that would change your mind. If you cannot say what would falsify the read, you are not analysing — you are narrating.
Sources
- primaryPublicly published central bank holdings and flow data