The consensus desk read is tidy: German wholesale prices continued climbing in September, pass-through to CPI is now a question of lag rather than direction, and EUR crosses should price a stickier ECB. The release lands inside a well-rehearsed story — energy base effects fading, intermediate goods repricing upward, services inflation refusing to normalize. The call writes itself. We have read it in three different notes this week, each citing the same wholesale print as evidence the inflation-risk thesis is intact. Hear the argument in its strongest form before we take it apart.
The Conventional Wisdom on German Wholesale Prints
The reasoning chain, laid out generously, runs like this. Wholesale prices in Germany are a leading indicator — producer and intermediate costs that have not yet reached the shelf. The German economy is the eurozone's largest input machine, so what reprices in Hamburg warehouses in September reprices in Berlin checkouts in December. If wholesale stays hot into Q4, the ECB cannot cut as aggressively as the futures curve had been pricing. Therefore EUR should rally. Therefore commercial accounts long EUR crosses are vindicated. Therefore every note writer who has been publishing the inflation-is-sticky thesis since July gets a victory lap in October.
The chain has institutional weight behind it. Bundesbank monthly reports have cited wholesale indices as early-warning signals since the 1970s. The BIS working papers tracking euro-area pass-through consistently frame German intermediate goods as a lead on the headline HICP by a window measured in months, not weeks. The Destatis release methodology has not materially changed in a decade. This is one of the cleaner series on the continent, and a desk that dismisses it entirely is dismissing a real signal.
So the trade writes itself in the sell-side format readers have been emailed all week. Buy EUR dips. Hold through the next ECB press conference. Wait for the pass-through to show up in the November and December HICP prints. Collect on the repricing of rate-cut expectations.
This is what every desk is writing right now. We are going to agree with most of it, concede the parts that are honest, and then show you the one place where the whole frame falls apart — and whose book you were reading when it did.
Why This Is Actually True
Start with the concession. German wholesale prices do lead the headline. Not perfectly, not every cycle, but the historical correlation is defensible on primary data. The 2021-2022 inflation surge showed up in wholesale prints months before it hit retail. The 2023-2024 disinflation telegraphed itself through the same channel. A desk that ignores the series is ignoring real information, and we would not pretend otherwise.
Second concession: pass-through from intermediate costs to final goods is not optional. The eurozone's internal pricing mechanics are what they are. If inputs cost more for long enough, outputs cost more. The lag varies — compressed in commodity-heavy sectors, stretched in services — but direction, over a long enough window, is settled. The honest debate is about magnitude and tempo, not about whether the mechanism exists.
Third concession: an ECB that cuts into a visible wholesale reacceleration looks political, and this ECB is unusually conscious of how it looks. Its own communication throughout 2025 has leaned on wholesale and PPI prints when defending its reluctance to cut aggressively. The staff projections, revised each quarter, treat the series as a top-tier input, not a sidebar. The September Economic Bulletin, which we have in front of us, does not pretend the number is irrelevant.
Fieldnote: three sell-side notes this week opened with the same chart — a twelve-month overlay of German PPI, wholesale, and HICP, drawn from Destatis. The chart is accurate. The authors are not wrong to publish it.
So when the consensus call says the September reread confirms the sticky-inflation thesis and tilts the ECB hawkish, every single link in that chain has a legitimate foundation. The lead-indicator property is real. The pass-through is real. The policymaker reaction function is real. If we stopped at the chart, the trade would be: get long EUR, wait, collect.
We are not going to stop at the chart.
Where It Breaks Down
The frame breaks at the question nobody writing these notes is asking: whose balance sheet benefits when every desk simultaneously writes the same thing?
Follow the commission structure. The retail FX industry runs on two broad revenue models and the split matters for which research reaches which inbox. The transparent-commission brokers — Pepperstone and IC Markets, in the operators we can cite here — charge an explicit per-lot fee and offer raw interbank spreads close to zero. The zero-commission brokers — XM and Exness — charge no visible fee but mark up the spread. On Exness's own disclosed figures, that gap is 1.0 pip average on the standard EUR/USD book versus 0.1 pip on the pro tier where the commission reappears in different form. On AvaTrade's published numbers, the standard book runs around 0.9 pips with no commission. FBS lists 0.7 on standard, 0.0 on the raw-spread tier. These are not editorial estimates. They are from the brokers' own public pricing pages.
Here is where the inflation-reread story stops being neutral. A wholesale print that generates one-way EUR directionality drives two very different revenue outcomes depending on which side of that model the broker sits.
For zero-commission books, strong directional conviction is poison on the headline minute. The spread markup compresses when flow goes one-way, because the broker has to warehouse the imbalance or hedge it in interbank at tighter prices than it is quoting to the retail account. The money is in two-sided churn — accounts that trade both ways through a session, each click paying a 0.6-0.9 pip markup embedded in the quote. A sticky-inflation story that tells every retail trader to hold long EUR for three weeks reduces that churn to a trickle. The incentive is to publish research that sounds confirmatory but leaves enough ambiguity to keep accounts flipping sides intraday.
For commission-model books, the opposite. The revenue line scales with volume regardless of direction. A high-volatility, directional EUR tape — exactly what a hawkish ECB reread would produce — maximizes lot count. Those research desks have no incentive to leave ambiguity. They want conviction, because conviction prints lots, and lots print commission.
Fieldnote: the three sell-side notes we read this week came from one commission-model shop and two independent macro desks. None from the zero-commission marketing teams, who published something else entirely — a "wait for confirmation" piece that read like it was written to delay flow commitment by a week.
Here is the primary-document cross-reference the consensus call is skating over. The ECB's September 2025 Economic Bulletin treats wholesale reacceleration as a signal the Governing Council should actively watch. The Bundesbank monthly report from the same window treats the same reacceleration as a signal already fading in forward-looking survey data not yet in the hard release. Both documents are operative. Both come out of the same system of central banks. They do not say the same thing. Read in sequence, they describe a Council watching a lagging confirmation while the regional authority is already seeing the next turn. That composition is not the chart in the three sell-side notes. The chart aggregates. The underlying does not.
The Rule I Use Instead
The rule, after a decade of reading these prints land into retail research cycles, is simple to state and harder to apply. Never trade a wholesale-driven inflation thesis without first identifying whose book the loudest research is written against. The macro reasoning can be entirely sound and the trade can still be a crowded entry that pays whoever executed before the note went out.
The operational version has three steps.
First, timestamp the research against the release. A note published 48 hours after a wholesale print that cites "stickiness" is not forecasting — it is positioning cover. The desks that mattered for the trade were already in it before the publication date. The retail-facing confirmation is the exit liquidity for the real money, not the entry. The timestamp alone will filter half the consensus chorus.
Second, read the primary release before any desk's interpretation of it. Destatis publishes the methodology. It is not opaque. The wholesale basket is weighted toward specific intermediate goods categories — fuels, metals, agricultural inputs — and a reread driven by one of those categories is a materially different signal than a reread spread across the basket. The consensus notes almost never specify which component drove the headline. The commission-model research desks sometimes do, in the longer-form client notes. The zero-commission marketing content, as a rule, never does, because specificity would narrow the directional story the content needs to leave ambiguous.
Third, check the revenue model behind the research channel you are reading. A commission-based broker's institutional research is not independent in any editorial sense, but its incentive is cleaner — volume regardless of direction. A spread-markup broker's content is written to maximize account churn, which means ambiguity when conviction would compress the markup and manufactured conviction when ambiguity would reduce clicks. Neither is honest in the way a reader hopes research will be honest. One is more legible once you know what you are looking at.
Fieldnote: the September release, read in isolation on the Destatis page before any note was emailed, shows two of the three components we track reaccelerating and one — the component most closely tied to EUR-denominated energy imports — flattening against August. That composition is not what the consensus overlay chart implies. The chart averages. The driver, in September, was narrow.
The rule does not generate a trade. It generates a filter. Before accepting the hawkish-ECB reread as the consensus desks are selling it, identify the commission model behind the research you are reading, timestamp the publication against the release, and read the component composition yourself from primary data. If, after that process, the thesis still holds its shape, then size the position appropriately. More often than readers expect, after that process the thesis looks less obvious than the chart made it look at first glance.
When the Old Rule Still Wins
Concede the limits. The incentive-tracing filter does not override the macro reality when the macro reality is overwhelming. The 2022 inflation shock did not care whose research desk was writing what — the signal was large enough to swamp any structural noise in the research channel, and every model, honest or marketing, pointed in the same direction. If a wholesale print reads at a magnitude that puts it in that regime — multiple standard deviations above the trend channel, broad-based across all basket components, confirmed in the parallel PPI series on the same day — then the trade is the trade and the incentive filter adds latency without any alpha.
The September print is not that regime. It is a modest reacceleration, concentrated in a subset of categories, inside a disinflation trajectory that remains, on every forward-looking survey measure the Bundesbank publishes, broadly intact. In the modest regime, the filter does the work it is designed to do. In the shock regime, it does not. We are not in the shock regime in October 2026. Readers who assume we are, because a twelve-month overlay chart looks superficially like the 2022 analog, are reading the wrong analog with the wrong research under their arm.