What runoff does mechanically
Quantitative tightening is usually run passively. Bonds held by the central bank mature, and instead of reinvesting the proceeds the holdings are allowed to roll off. The asset leaves the central bank balance sheet, and because every balance sheet balances, a liability leaves too. In practice the liability that shrinks is bank reserves or a related cash balance.
There is a second effect on the private side. Debt that the central bank was holding has to be held by somebody else instead, which means private investors absorb more duration than before. Absorbing duration requires compensation, so term premium tends to rise. These are two genuinely different channels. One is about the quantity of settlement cash in the system. The other is about who holds interest rate risk and at what price. Commentary routinely merges them into a single idea called liquidity, which is where most of the confusion begins.
Reserves are a real constraint, not an abstraction
Bank reserves are the balances commercial banks hold at the central bank, and they are what banks use to settle payments between each other. Banks want to hold a buffer of them for liquidity regulation, for intraday payment needs and for simple prudence. That demand is genuine and it is not fixed. It shifts with regulation, with the size of bank balance sheets and with how nervous institutions feel.
The critical point is that nobody knows the level of demand in advance. You discover it by crossing it. Conditions move through a rough sequence from abundant, where the quantity of reserves is far above what anybody needs and marginal changes do nothing, to ample, where it is comfortably sufficient, to scarce, where overnight funding rates start to firm because somebody is genuinely short. The transitions are not announced and they are only obvious in hindsight.
The rate channel and the liquidity channel are not the same
The policy rate sets the carry cost of holding anything that pays nothing. That is the channel most gold commentary already understands, and it is covered in interest rates and gold. Balance sheet policy works differently. It operates on the quantity of cash in the system and on the amount of interest rate risk the private sector has to hold.
Those two can move in opposite directions. A central bank can cut its policy rate while still allowing holdings to run off, which loosens the carry channel and tightens the quantity channel at the same time. Markets then get a muddled signal, and simple narratives fail. Keeping the two separate in your head is the single most useful discipline here, because it immediately explains sessions where the metal behaves in a way that seems inconsistent with the rate decision just announced.
Where the liquidity actually went
This is the detail that defeats most liquidity overlay charts. Central bank liabilities are not only reserves. There is typically an overnight facility where money market funds park cash, and there is a government cash account that fills with tax receipts and bond proceeds and empties with spending. Both act as buffers.
When runoff shrinks the asset side, the offsetting fall can come out of any of those. If the overnight facility is draining, the balance sheet can shrink substantially with bank reserves barely moving, and funding conditions remain comfortable. Once that buffer is exhausted, further runoff of the same size starts coming directly out of reserves, and the system becomes sensitive very quickly. Identical headline numbers therefore mean completely different things depending on where the buffers stand. Anyone drawing conclusions from the total alone is reading the wrong line.
The two directions gold feels this
The first direction is supportive and slow. A rising term premium and persistent fiscal supply are part of a broader story about the burden of government debt, which is one of the structural reasons reserve managers and long horizon investors hold metal at all. That bid does not respond to weekly changes in a balance sheet number.
The second direction is the opposite and much faster. When reserves become genuinely scarce, funding gets dearer for anyone running leverage. Financing a position costs more, levered holders reduce, and if the scarcity becomes acute enough, cash wins over everything and liquid assets get sold to raise it. In that window the metal trades poorly regardless of the structural argument. The figure shows the regime bands and why their boundaries cannot be pinned down in advance.
Why liquidity charts mislead
Overlaying a liquidity proxy on a gold chart is one of the most popular exercises in macro commentary and one of the least reliable. The problems compound.
- The proxy is usually assembled from a few balance sheet lines chosen because the resulting line happened to fit. Change the recipe and the fit changes.
- Reserve demand is unobservable and shifts, so there is no stable threshold for the proxy to be compared against.
- The relationship is regime dependent. Liquidity matters enormously near scarcity and barely at all when reserves are abundant, so a single correlation measured across both regimes is meaningless.
- Balance sheet data is weekly and backward looking, while the chart moves continuously.
None of that means the channel is not real. It means the channel is real and the popular measurement of it is weak, which is a different and more annoying situation.
A sane way to track it
Watch the symptoms rather than the cause. Scarcity shows up first in overnight funding markets, as secured rates pressing against or above the administered ceiling, and as persistent rather than one off prints. It shows up next in the behaviour of the long end, where term premium does the work. Neither requires a homemade liquidity index.
Then let the tape arbitrate. If funding is comfortable, balance sheet policy is background noise for a gold position and the rate and currency channels dominate. If funding is tightening, raise your guard, expect poorer fills and reduce size before the chart tells you to. The context sits alongside the longer yield picture in treasury yields and gold, while entries still come from structure on the live chart. Treating the balance sheet as background until funding symptoms appear keeps your attention on the part of this channel that actually bites.
FAQ
Is quantitative tightening simply bearish for gold?
Not reliably. It raises term premium, which is a headwind through real returns, while also contributing to a structural story about debt that some investors read as supportive. The sharper effect comes if reserves become scarce enough to tighten funding, and that effect is short lived and two sided.
What exactly is a bank reserve?
It is a balance a commercial bank holds at the central bank, used to settle payments with other banks. Banks hold a buffer for regulatory and practical reasons. Reserves cannot be lent out of the banking system as a whole, which is why the useful question is the size of the buffer relative to demand.
Why does the overnight facility drain before reserves?
Because cash parked there is the most footloose money in the system. As other short term instruments become relatively more attractive, money moves out of the facility rather than out of bank settlement balances. That absorbs the drain for a while, which is why headline shrinkage can proceed without any funding effect.
Does runoff push long yields up?
It removes a large and price insensitive buyer, so private investors must hold more interest rate risk and generally require more compensation. That tends to lift term premium. The effect is one input among many, and issuance patterns, foreign demand and growth expectations can easily overwhelm it in the short run.
How quickly does any of this show up in gold?
The structural channel works over quarters and is invisible day to day. The funding channel can show up within a session once stress is genuine, usually as poorer liquidity and sharper moves rather than as a clean directional trend. Neither offers a timing signal on its own.
ⓘ See these ideas on real price: open the free XAUUSD live chart.