Let me concede something upfront. When TD Securities publishes a note calling the United States "sideways growth with sticky inflation," the phrase does real work. It compresses four quarters of confusing data into a shape a portfolio manager can hold in one hand. That is a genuine skill and I do not want to pretend otherwise. But the phrase is also a spread — a specific gap between two moving numbers — and spreads have layers. The interbank layer. The desk markup. The liquidity premium at 3 a.m. Tokyo time. Once you decompose it, the call becomes something more interesting than a headline.
The desk that writes this publication tends to distrust two-clause macro descriptions on reflex. Not because they are wrong. Because the compression is where the interesting information lives, and the compression is the first thing summarized away. So we want to do something specific with this note: treat it as if it were a bid-ask quote on a trading screen and take it apart the way a market-making desk would.
There is a reason we care about doing this from an Asia Pacific vantage point in particular. The Tokyo and Singapore books that price dollar risk into the London open are looking at a different set of receipts than the New York strategist who wrote the note. They see the flow first. They see the hedging demand from Japanese life insurers before the note's readers do. And what a "sideways growth with sticky inflation" call means in those books is not what it means in the coverage summary.
Let Me Concede the TD Securities Call Before I Take It Apart
Here is what the call gets right, and I want to lead with this because if you skip this section you will read the rest of the piece as a hit job and it is not one. A "sideways growth" description of the United States in late 2026 is defensible if what you mean is that the point estimates on quarterly real GDP have been running inside a narrow corridor for long enough that mean-reversion arguments have stopped feeling like arguments and started feeling like observations. That is a real thing. It is what happens when a large economy digests a rate cycle without cracking. The 1994-1995 Greenspan soft landing had a similar texture on the growth side, and desks that lived through that period remember it as a market where the story was quiet and the flow was not.
"Sticky inflation" has similar defensible content. What TD Securities is doing there, if we read the phrase generously, is refusing to celebrate a headline CPI print that has come down while services and shelter components have not. That refusal is analytically correct. Anyone who lived through the 1994 or the 2005 soft-landing calls remembers that the surprise never comes from the component that everyone is watching. It comes from the one that has been quietly ratcheting.
But — and here is where the concession ends — those two clauses are not independent. They are describing the same economy at the same time, and if you write them as if they were two separate observations you have produced a note that a portfolio manager can carry to a committee meeting, not a note that a trading desk can act on. And this is where the desks that price dollar risk into the Asian session start to squint.
The tell is this. If growth is genuinely sideways and inflation is genuinely sticky, then the real policy rate — the nominal policy rate minus the inflation the market actually expects — is not stable. It is moving. Slowly, but moving. A Tokyo book that funds a US-dollar exposure at the overnight rate cares about that spread with a precision that a strategist writing a quarterly note does not need to. The desk needs the derivative. The note gives the level.
You can hold both clauses of the TD Securities call as true and still conclude that the market underneath them is being repriced every session, in small increments, by exactly the kind of participant whose flow does not show up in the note's data tables. That is not a criticism of the note. It is a description of what a spread is.
The Spread Between Growth and Inflation Has Layers, and Tokyo Prices Them First
OK, here is the part I actually wanted to write. If you have ever wondered why dollar pricing behaves differently between 10 p.m. Tokyo and 10 a.m. New York on the same day, it is not random and it is not sentiment. It is a very specific thing about which books are open at which time, and I love this detail, so let me walk through it with numbers.
Consider a EUR/USD standard-account spread as it appears on a retail platform. The grounding for this piece lists AvaTrade at an average of 0.9 pips on the standard book, Exness at 1.0 pips on standard and 0.1 pips on its Pro tier, FBS at 0.7 pips standard and 0.0 pips on the raw ECN tier, and HF Markets at 1.2 pips standard and 0.0 pips on the Pro side. Those numbers are the observable surface. They are also what almost every article on this topic stops with.
Now let us decompose. Take the 1.0-pip Exness standard number and read it as a sum. The interbank leg — the raw bid-ask on the top-of-book at a tier-one prime broker feed — is running somewhere around 0.1 pip on major-hour EUR/USD, which is exactly what the Pro tier discloses. That means the standard-tier markup is 1.0 minus 0.1, or 0.9 pip. Nine-tenths of the total spread the retail client is paying is not the cost of the market. It is the cost of the retail order flow being commingled with other retail order flow inside the broker's B-book risk system. That is what a "standard" spread is.
Apply the same decomposition to FBS. Standard 0.7 pip, Pro 0.0 pip. The markup is 0.7 pip. To HF Markets: standard 1.2, Pro 0.0, markup 1.2. To AvaTrade: standard 0.9 with no explicit raw-spread account disclosed, so the markup is unresolvable from the observable numbers alone — which is itself a piece of information about the broker's model.
Now stack this on top of a macro call. TD Securities telling you the United States is sideways-plus-sticky is telling you, in trading terms, that the interbank layer of the dollar pair is going to compress. Low realized vol on a stable rate spread produces tight top-of-book. That 0.1 pip interbank number will hold or narrow. What it does not tell you is that the retail-facing 1.0 pip will hold. The markup layer widens the moment a headline hits, because the broker's risk desk widens it defensively, and it widens hardest during the Asian session, when the depth of book on EUR/USD is at its thinnest.
Let me finish the math. Take a trader running a 5-lot position — 500,000 EUR notional — through a standard account at the 1.0-pip mid. One pip on 500,000 EUR/USD is 50 dollars. Round trip, that is 100 dollars in spread cost per rotation. If they rotate the position four times during the Asian session hoping to trade a headline reaction to a US inflation surprise, that is 400 dollars in transactional friction against a headline they are reading two hours after the New York desks have already faded it. Now imagine the broker widens the standard spread from 1.0 to 1.6 during the print window — a routine widening for Asian-hours EUR/USD around a US data release. The rotation cost jumps from 400 to 640 dollars. That 60 percent increase is not on the trader's screen anywhere until the fills come back.
This is what TD Securities' phrase is priced into. Not the interbank leg — that responds to the note within seconds of publication and does not care about retail. The retail markup. The Asian-hours liquidity thinning. The desk-defensive widening in the four minutes on either side of a CPI print. Every one of those layers moves independently of the headline, and every one of them shows up in what the trader actually pays.
There is a reason Saxo Bank APAC, Interactive Brokers Asia, OANDA Asia and IG Group Asia disclose their execution statistics differently than their European parents do. The Asian session is where the layering is most visible and least discussed. Whenever a note like the TD Securities call lands overnight New York time, the first books to trade it are the Tokyo and Singapore books, and the first spread to widen is the retail-facing one on the platforms serving that region. By the time London opens, the reprice is done, and the note reads like it was already in the price. It was — for the desks in the right time zone.
What This Desk Would Watch Instead of the Headline Print
The reflex when a US macro call lands is to watch the next CPI or the next payrolls print for confirmation. That is what the note is written to be tested against, and I understand the impulse. But if you have followed the argument this far, you already know what I am going to say. The confirmation is not in the headline. It is in the layers.
The first thing this desk would watch is the front-end swap curve on the two-hour window that brackets each major US release. Not the level. The shape. A "sideways growth, sticky inflation" world produces a curve that flattens on the short end without moving the long end — because the market is repricing when policy responds without changing what policy eventually settles at. That is a very specific footprint. The 2004-2006 tightening cycle left the same shape on the curve, and desks that were on the JPY carry trade during those years remember watching for it because it was the earliest signal that the yen funding leg was about to become dangerous.
The second thing to watch is the Asian-hours EUR/USD depth of book. Not the spread — the depth. If the top-three-levels of the book are showing less quoted size than the six-week rolling average, the retail-facing markup on standard accounts will follow. That is the pass-through mechanism I described in the math section. Depth thins, markup widens, and the trader pays the difference without ever seeing the interbank leg move. The MAS Singapore wholesale market framework established in 2008 is the regulatory scaffolding under which Singapore-based prime brokers post those quotes, and the framework was written specifically for the visibility that Asian-hours pricing needs. The framework works. What defeats it is the retail platform layer that sits above it.
The third thing to watch is not a number at all. It is the language of the next note the same desk publishes. "Sideways growth with sticky inflation" is a phrase that can hold for two quarters. It cannot hold for four. Either the growth clause moves — up or down — or the inflation clause resolves. When the same desk publishes the note that changes one of those clauses without acknowledging that it changed, that is the moment the earlier call was doing more compression than analysis. It happens. It is not a scandal when it happens. But it is the moment where you learn how much of the earlier call was signal and how much of it was portfolio-manager-friendly packaging.
There is a version of this piece that treats the TD Securities call as a straw man and pulls the whole thing apart. That is not the piece I wanted to write. The call is defensible. The compression is the point of the call, and the compression is what makes it useful to the people who need it in that form. What I wanted to do was show what a desk that lives one time zone ahead of the note's audience actually sees when the note lands, and how the layers of the spread underneath it get repriced before the note's readers finish their coffee.
This piece started as a decomposition of a single macro call and turned into an argument about the retail spread. That was not the plan when I sat down. Once I started pulling on the interbank-versus-markup thread, it became clear that the reader who cares about the TD Securities note in the way it wants to be read — as a directional view — is not the reader who is going to be most affected by whether it is right. The reader who pays for the layers underneath is the one who needs the anatomy. So the anatomy is what the piece became.
FAQ
What does "sideways growth with sticky inflation" actually mean in trading terms?
It describes a regime where real GDP prints stay inside a narrow corridor while services and shelter components of CPI refuse to compress. For a trading desk, the operational implication is that the real policy rate — nominal minus expected inflation — drifts even when the headline policy rate holds. That drift is what gets priced into the front end of the swap curve, and it is the layer that repeats through into overnight funding costs for dollar-funded positions in the Asian session.
Why does this desk keep talking about the Asian session specifically?
Because the Tokyo and Singapore books are open when overnight US notes land, and they price the reaction first. By the time London opens, the reprice is done. Retail spreads on EUR/USD widen more during Asian hours than during New York hours because the depth of book is thinner and the retail-facing markup responds defensively to headlines that hit outside the deep-liquidity window. The mechanics are the same everywhere. The visibility is highest in Asia.
How do I decompose a broker's advertised spread on my own?
Take the standard-account average spread and subtract the raw-spread or Pro-tier equivalent. The difference is the markup. Exness at 1.0 pip standard versus 0.1 pip Pro gives a 0.9 pip markup. FBS at 0.7 standard versus 0.0 Pro gives 0.7. HF Markets at 1.2 versus 0.0 gives 1.2. When the broker does not disclose a raw-spread equivalent — AvaTrade at 0.9 with no Pro-tier comparison in the grounding — the decomposition is unresolvable from public numbers, which is itself information about the broker's model.
Is the TD Securities call wrong?
No, and I said this in the concession. The call is defensible and the compression it performs is what makes it useful for the audience it is written for. What the piece argues is that the call is a summary of a spread rather than an observation of a variable, and spreads have layers that the summary does not carry. Whether the call is "right" in the directional sense is a separate question from whether it is being priced correctly at the retail-facing layer.
Why should an APAC-based reader care about a note written from a North American desk?
Because the note's audience is trading in a time zone where the flow it describes has already moved through the book by the time they read it. An APAC reader is one time zone ahead of the reprice, which means the reprice is happening on their screens before it lands as a headline for the note's intended audience. That asymmetry is not exploitable in an easy way, but it changes what "reacting to the note" means.
What regulatory frameworks matter for the Asian-hours execution described here?
The MAS Singapore wholesale market framework of 2008 governs the prime-broker liquidity layer that sits under Singapore-based retail platforms. The HKMA linked-rate regime, in place since 1983, shapes Hong Kong dollar funding for USD-referenced positions. Japan's JFSA framework under the 2005 FX law caps retail leverage in a way that changes the composition of Tokyo-session flow. Each of these frameworks is worth reading if the reader is trying to understand why Asian-session pricing looks the way it does.
What did this piece deliberately not cover?
Three things. It did not cover the tax treatment of spread costs versus commission costs under the various APAC jurisdictions, because the treatment differs by country and this desk is not qualified across all of them. It did not cover the option-implied volatility surface for USD pairs during the print window, which is a separate and larger topic. And it did not cover the specific portfolio construction implications of a "sticky inflation" call for a fixed-income book, which is where the note's own audience would want the argument to go.