Most desks read the People's Bank of China's move to daily overnight reverse repos through November 2 as a straightforward follow-through to its October policy rate signal — a plumbing adjustment that confirms the direction of travel. We would push back on that framing. The daily overnight window is not the second step after a rate signal; it is a separate instrument, operating on a different horizon, aimed at a different audience inside the interbank system. Treating it as confirmation flattens what the PBOC is actually doing, and it leaves the reader unprepared for what the November 2 expiry will — and will not — mean.
The Consensus Reading of the PBOC's Daily Overnight Window
Walk into any sell-side morning call this week and you will hear a version of the same story. The PBOC signalled on the policy rate in October. It then opened a daily overnight reverse repo facility through November 2. Ergo, the overnight window is the operational echo of the rate decision — the way the central bank seeds the signal into the funding curve so that short-end rates move toward the policy intention. The arrow runs cleanly from policy announcement to open-market execution to money-market pricing. One narrative. One direction. One continuous act.
The story is tidy, and tidiness is what sell-side notes reward. A client gets a two-line summary that stitches two headlines together, the strategist gets a view that cannot easily be falsified before November 2, and the trader gets a bias that lines up with the loudest voice on the Reuters chat. If you have been in markets for less than three years, this is probably the frame you absorbed on day one — the central bank talks, then the central bank acts, then the curve prices it in, in that order.
I understand the appeal. I used to read it the same way. When you are new to this — and most of you reading this will have come into markets after 2020 — you look for continuity because continuity is how you tell yourself you are learning. Everything feeds into one story. The problem is that the PBOC is not writing one story. It is running two different machines at two different speeds, and it has told you so, in print, if you are willing to go find the documents.
Why This Is Actually True
Here is where I want to be fair, because the consensus is not wrong about everything, and if I pretend it is I lose you. There is a legitimate reason to connect the October rate signal and the daily overnight window. Both are statements the PBOC chose to make inside the same four-week window. Both land on front-end liquidity. And in a central-bank communication environment where silence is itself a signal, the sheer fact that the Open Market Operations desk is now printing a line in the morning — every morning, through November 2 — matters. If the central bank wanted the market to see dissonance, it could have paused.
So the first thing the consensus gets right is the base rate observation: when a major central bank layers two actions on top of each other in the same monetary quarter, the default prior is that they rhyme. In the Federal Reserve's post-2008 playbook, forward guidance and standing facilities moved in lockstep for exactly this reason — the staff did not want the standing facility to appear to contradict the FOMC statement. The PBOC is a different institution with a different toolkit, but it reads the Fed closely enough that it understands the cost of accidentally dissonant plumbing moves. If the overnight window were cutting against the October signal, we would expect that to show up in the Q3 monetary policy report, and it would show up as a correction, not as a continuation.
Second, the consensus is right that the daily cadence itself is a message. Weekly or ad-hoc operations are more common; daily is a posture. It tells the primary dealer network that the OMO desk is standing there, every morning, before the open, as a visible counterparty. That alone narrows the right tail of overnight funding spikes, which is useful whether or not the policy rate moved. Front-end volatility compression is a legitimate benefit, and if the consensus story stopped there, I would nod along.
But here is what that framing misses — and misses in a way that will cost you money on November 3 if you have been trading it as a continuation.
Where It Breaks Down
The consensus reading breaks down in two places, and the second one is the one I want you to carry away.
The first break is horizon. The October policy rate signal is a stock variable in the market's head — it resets the long-run anchor for the entire RMB curve, out past the one-year LPR, out into the five-year, out into the mortgage-reset calendar for 2027. The daily overnight reverse repo is a flow variable. It injects liquidity for one business day. If you roll it every day from the start of this window to November 2, you have perhaps fifteen or sixteen business days of overlapping daily tranches, and the gross headline numbers that will scroll across your terminal are going to look enormous while the net stock of liquidity in the system stays modest. Those are not the same instrument wearing different hats. They are two different instruments, and the consensus story blurs them because the blur is comfortable.
Let me show you the arithmetic so you can reproduce it. Suppose the OMO desk sets a daily overnight tranche at an illustrative CNY 200bn. Day one: gross injection 200, net stock 200. Day two: another 200 goes out, but day one's 200 matures the same morning, so net stock stays at 200 while gross injections tick to 400 cumulative. Day five: cumulative gross is 1,000; net stock is still 200. Day ten: cumulative gross is 2,000; net stock, 200. Day fifteen — call it the full run to November 2 — cumulative gross is 3,000; net stock is 200. A reader scanning only the headline injection number will see a chart that looks like liquidity flooding into the system. A reader who tracks the net has not seen the stock move at all. One headline. Two completely different stories. The desk that trades off the gross number is being set up to be wrong on November 3 when the window closes and the gross injection line goes to zero — because the net never moved to begin with, so there is nothing to withdraw.
The second break is audience. The policy rate signal is aimed outward — at corporates planning capex, at mortgage borrowers, at offshore CNH watchers. The daily overnight window is aimed inward, at the primary dealer network and the handful of large joint-stock banks whose repo books carry the plumbing of the interbank market. These are two different conversations the PBOC is having, in two different rooms, and treating them as one conversation is how you end up positioning on November 3 as if the window closing was a hawkish signal. It is not a signal at all. It is the end of a plumbing operation whose audience was always the dealer desk, not you.
The Rule I Use Instead
Here is the rule, and it works for every central bank, not just the PBOC. Separate the signal from the plumbing before you position.
A signal is an announcement whose intended function is to reset the market's prior about the future path of a policy variable. The October rate action is a signal. It is designed to be read, interpreted, and repriced into forward curves by anyone with a Bloomberg terminal, in Shanghai or London or São Paulo.
Plumbing is an operation whose intended function is to control the mechanics of the money market in a specific window. The daily overnight reverse repo is plumbing. It is designed to narrow the realised variance of front-end rates, so that the dealer network can clear overnight funding without the kind of spikes that have historically shown up in Chinese quarter-end reporting dates. Plumbing can be running at the same time as a signal, in the same direction as a signal, and completely independent of a signal — and often is all three at once, in different proportions, which is why the consensus story feels right even when it is wrong.
The two sources you need to read if you want to tell them apart are both in the public record. The PBOC's quarterly monetary policy executive report is where the signal is explained — read the "Next Stage Policy Ideas" section and you will find the central bank arguing explicitly for the stock variable. The PBOC's daily Open Market Operations announcement is where the plumbing is logged — a one-paragraph release, usually before 10:00 Beijing time, specifying the day's operation size and tenor. If these two documents are saying the same thing, you have continuity. If one is talking about the policy rate anchor and the other is talking about reserve maintenance smoothing — which is often the honest text of the OMO note — then you have two different things, running on two different horizons, and the trade is to position the signal and ignore the plumbing cadence.
In the specific case of the current window: read the OMO announcements that bracketed the policy rate decision, and read the executive report that framed it. If the executive report emphasises the policy rate as the lead instrument and the OMO notes talk about maintaining "reasonable and ample" liquidity through quarter-end — that is not confirmation. That is two separate statements that happen to be in the same news cycle. Position accordingly.
When the Old Rule Still Wins
I have to concede the limits of my own frame. The signal-vs-plumbing rule works when the central bank runs a disciplined separation between its policy desk and its OMO desk — when the plumbing operations are small relative to the balance sheet, incremental, and documented in neutral language. In those environments, you can treat the two as independent and you will be right most of the time.
The rule fails, and the consensus story wins, when the central bank is deliberately using the plumbing as communication. The Bank of England's 2022 gilt market operations are the textbook case: an ostensibly technical facility, scaled large enough and announced loud enough that it was functionally a signal about rate expectations. If the PBOC starts scaling the daily overnight window dramatically — if the per-day tranche size moves well above its recent range, or if the window gets extended past November 2 with explicit new language in the OMO notes — then the plumbing has become the signal. In that regime, the consensus reading is correct and my rule is wrong. Watch the per-day size and the language. If both stay inside their normal range through November 2, trust the separation. If either breaks out, re-read the window as a signal and reprice.
This piece does not cover the parallel question of how the daily overnight window interacts with the one-year MLF renewal calendar, which is where the stock variable actually gets set — that is a separate argument that deserves its own article. It does not cover the offshore CNH funding transmission, which runs on a different plumbing network entirely. And it does not cover the question of whether the policy rate signal itself was correctly priced in October — reasonable desks disagree, and the resolution is not inside the scope of a piece about reverse repos.
FAQ
What is the PBOC's daily overnight reverse repo window and why does it run to November 2?
The daily overnight reverse repo is a one-business-day liquidity injection run by the PBOC's Open Market Operations desk, settled against collateral, that matures the following morning. Running it on a daily cadence through November 2 means the OMO desk is committing to appear in the market every morning in that window, which compresses front-end funding variance. The end date is a cadence commitment, not a maturity cliff — the net stock of liquidity does not drop meaningfully when the window closes.
Does the daily overnight window confirm the direction of the October policy rate signal?
Not in the way the consensus suggests. The policy rate signal operates on the long-run anchor of the RMB curve; the daily overnight window operates on realised front-end variance over roughly fifteen business days. They can rhyme, and in the current case they do not obviously contradict, but confirmation would require the OMO notes to use signal-grade language. Plumbing cadence by itself is not confirmation — it is a parallel operation aimed at the dealer network.
What happens to RMB liquidity on November 3 when the window closes?
Mechanically, far less than headline injection numbers would suggest. Because each daily overnight tranche matures the next morning, the net stock of liquidity added by the window stays roughly flat throughout — only the gross cumulative injection grows. When the daily cadence stops, the gross number stops growing and net stock is unchanged. If you have been positioning off the cumulative gross figure, you have been trading an illusion.
How do I actually tell a signal apart from a plumbing operation?
Read the two primary sources side by side. The PBOC's quarterly monetary policy executive report explains the signal and the policy rate anchor in explicit forward-looking language. The daily OMO announcement explains the plumbing operation in neutral maintenance language — size, tenor, phrases like "reasonable and ample liquidity." When the executive report leads with the policy rate and the OMO notes read as technical maintenance, you are looking at two separate instruments, not one continuous act.
When should I trust the consensus "it's all one signal" reading instead?
When the plumbing operation is scaled large enough or announced loud enough that it is functionally communicating a policy stance. The Bank of England's 2022 gilt market facility is the archetype — nominally technical, operationally a signal. If the PBOC's per-day tranche size breaks sharply above its recent range, or if the window is extended past November 2 with new emphatic language, the plumbing has crossed into signal territory and the consensus reading becomes the right one.
Are there offshore implications for CNH funding from this window?
The onshore daily overnight window operates on the mainland interbank system and the primary dealer network that clears it; the CNH funding market runs on a different plumbing network with its own liquidity dynamics. There is a transmission mechanism, but it is slower and noisier than onshore front-end rates. Any direct read-across from daily OMO cadence to CNH overnight pricing should be treated with caution — it is a separate article's worth of mechanics.
What should a newer trader actually watch between now and November 2?
Three things. The per-day tranche size relative to its recent range — breakouts matter, steady readings do not. The language in the daily OMO announcement — a shift from "reasonable and ample" to anything more directional is the real signal. And the net versus gross liquidity picture — track the maturities, not the headlines. If you only have time for one, watch the per-day size against its recent envelope.