Next Tuesday the FOMC minutes land, and Governor Bowman's "cautious, data-dependent pause" phrasing will be reread by every desk that was long dollars into the hike. We are writing this on a weekday afternoon with twenty-four hours of lead time, because the spread you pay between 1:58 PM and 2:04 PM ET on release day is not the spread your broker advertises on its homepage. That gap — between the quoted 0.1 pip raw and the actual fill — is the subject here. It has a history. It has operators. It has a cost nobody prints on the dashboard.

We are not writing a Fed preview. There are two thousand of those. We are writing about what happens to your execution during the six minutes Bowman's language gets reread. The two subjects turn out to be the same subject, once you look at where the money actually leaves the account.

The Pause Language Is a Volatility Instruction, Not a Dovish Signal

The word "cautious" does more work than traders give it credit for. A dovish pause tells the desk the next move is a cut. A hawkish pause tells the desk the next move is a hike. A cautious, data-dependent pause tells the desk nothing about direction and everything about dispersion. It is an instruction to widen the distribution of outcomes around the next two prints — CPI, nonfarm payrolls, PCE. That is a volatility signal, not a rate signal.

Volatility desks price this correctly. Retail platforms price it a different way, which we will get to.

Here is the texture most previews leave out. When a governor adds the modifier "data-dependent" to a pause statement that already contained the word "cautious," they are telling the market that the committee's reaction function has become more sensitive to the next print, not less. Sensitivity is the parameter that gets widened. The policy rate stays flat; the implied volatility on front-end rates futures lifts. This is a known pattern, documented across the Bernanke-Yellen transition, the 2016 Yellen pause, the 2019 "mid-cycle adjustment" Powell framing. Every one of those episodes saw EUR/USD realized volatility rise in the week after the pause language, even though the policy rate did not move.

The historical reconstruction worth sitting with is not Volcker or Greenspan. It is October 2008.

In early October 2008 the FOMC was cutting into a crisis. The language that moved markets was not the 50bp cut itself. It was the single sentence in the statement that said the committee stood ready to "act as needed" — a phrase the desks reread for four hours after the release. Interdealer EUR/USD spreads during those four hours, on the then-dominant ECN platforms EBS and Reuters Matching, widened from the typical 1-2 pip institutional quote to something closer to 5-7 pips during the volatile windows. Not because liquidity vanished entirely. Because the three or four tier-one makers who were still quoting required a cushion to warehouse the risk. That cushion is called the spread. In October 2008 the cushion was named crisis. In 2026, the equivalent cushion is called "normal market conditions" and nobody writes about it.

The lesson for next Tuesday is boring and expensive at the same time. Bowman's phrasing does not tell you whether to be long or short dollars. It tells you that whatever position you are carrying, the cost of adjusting it between 2:00 PM and 2:06 PM ET will not be the cost advertised on your broker's spread page. Behave accordingly.

Free Download
Broker Red-Flags Checklist (PDF)
15 red flags that expose a bad broker in minutes — plus the 2 brokers that pass all 15. No fluff, print it.

Spread Markup During FOMC Minutes Is the Industry's Oldest Unadvertised Tax

This is where we have to say some things the industry prefers unsaid.

The spread on a retail forex platform is not one number. It is three. There is the raw interbank spread, which is what prime brokers show tier-one makers. There is the operator's markup, which is what the broker adds. And there is the execution spread, which is what you actually pay including slippage. On a quiet Wednesday morning these three numbers are close enough that nobody notices the difference. During the six minutes around an FOMC release they are not close.

The historical arc here matters, and it is the specific territory this desk covers. Before 2001, retail forex barely existed and institutional spreads on EUR/USD routinely traded at 3-5 pips on EBS during normal conditions, wider in London afternoon illiquid windows. The emergence of ECN infrastructure through the early 2000s compressed those institutional spreads to 1-2 pips by the mid-2000s. The retail CFD compression came later and faster. By 2015, operators advertising "raw" spreads under commission-plus-raw models — IC Markets Raw, Pepperstone Razor, FXCM Active Trader, Tickmill Pro — were quoting EUR/USD at 0.1 to 0.3 pips during London overlap. The quoted number collapsed by roughly 95% in fifteen years.

What did not collapse is the markup during news.

The industry's quiet dirt — and we mean this literally; it is not a conspiracy, it is the mechanics of how market makers survive — is that the compression was in the headline number only. During FOMC minutes release windows, during NFP, during ECB press conference first-questions, the spread on EUR/USD on retail platforms routinely widens to 3-15 pips. Not because the raw interbank spread widens by that much. Because the operator's markup, which averaged perhaps 0.1 to 0.5 pips during normal conditions, lifts by a factor of ten to twenty.

The mechanism has a name on institutional desks: A-book / B-book routing. On the A-book, the broker passes your order to a liquidity provider and takes a commission. On the B-book, the broker takes the other side and warehouses the risk. During quiet conditions the B-book is profitable because retail traders systematically lose money and the broker keeps the loss. During the six minutes around Bowman's language getting reread, the B-book is not profitable because the directional move eats the house. The response, documented across the industry since the mid-2010s, is to widen the quoted spread so that even if your stop gets hit the broker captures the markup. This is not stop-hunting in the conspiratorial sense. It is risk pricing in the actuarial sense. The outcome for you is identical.

Two primary document types describe this, and they describe it differently, which is the point. ESMA's August 2018 product intervention measures on CFDs, finalized across the EU, imposed leverage caps and negative balance protection — but notably did not impose spread-markup disclosure requirements. ASIC's October 2020 product intervention order on retail OTC derivatives in Australia mirrored the leverage caps but, in paragraph after paragraph of the explanatory statement, dwelt on execution quality and the "material widening of spreads at times of high volatility." Both documents are operative in 2026. The ESMA framework treats spread as a product-design feature; the ASIC framework treats spread widening as a conduct issue. One operator can be fully compliant under ESMA and still fall short of the ASIC conduct standard. This is not legal paradox. It is regulatory divergence that the retail reader is almost never shown.

The YouTube thumbnail said "zero spread, instant execution." The YouTube video, when we watched it carefully, showed a demo account trading at 3:30 AM London time on a Sunday. We are not making this up. The demo account was quoting a 0.0 pip EUR/USD spread on an untraded instrument during a session when the interbank market is closed. The quote is technically true. It is also structurally meaningless. Zero divided by zero is not a spread; it is an absence of market. The entire genre of "raw spread" marketing elides the distinction between the quoted number during quiet hours and the executed number during news.

The practical response is a routine that most retail traders refuse to adopt because it feels paranoid. Before Tuesday, if you are carrying a dollar position, decide now whether you plan to hold through the release or flatten before it. If you plan to hold, size your position for a 15-pip adverse move that has nothing to do with the fundamentals and everything to do with your fill. If you plan to flatten, do it before 1:55 PM ET, not at 1:59. The last four minutes before an FOMC release are when operator markups begin to lift pre-emptively, documented in execution-quality reports that the better-regulated operators publish quarterly and that nobody reads.

The Only Thing That Would Change This Reading Is a Register of Execution Logs

We would reverse the position taken in this piece if the FCA, ASIC, or CySEC — the regulators under which most of the operators named above hold tier-one licenses — published a mandatory register of execution logs for retail CFD trades during the six-minute windows surrounding scheduled macro releases. Specifically: broker-level reporting of average quoted spread, average executed spread, slippage distribution by percentile, and A-book versus B-book routing ratios, published within thirty days of each FOMC, ECB, BOE, and BOJ release.

That register does not exist. We have asked why for several years in various forms. The honest answer, from people who work on the industry side of the conversation, is that the data would embarrass most operators and would materially reduce retail trading activity around scheduled news events. The second honest answer is that regulators, with limited enforcement bandwidth, have prioritized leverage caps and negative balance protection — both of which are concrete, measurable, and politically defensible — over execution-quality disclosure, which is technically complex and commercially contested.

The counterfactual version of the forex industry — the one where that register exists — would not look dramatically different from the outside. Spreads would still widen around news. B-book routing would still happen. The difference is that the reader before next Tuesday's FOMC minutes would be able to pull up a document and see, broker by broker, what the median EUR/USD spread was between 2:00 PM and 2:06 PM ET across the last twelve FOMC releases. That number would do more to shape retail execution behavior than any amount of educational content about Fed policy.

Until that register exists, the argument in this piece holds. The pause language from Bowman is a volatility instruction, the retail spread architecture turns volatility into a hidden tax, and the only observable response for the individual trader is to assume the spread you pay next Tuesday afternoon will not be the spread quoted on the operator's homepage. We would like to be wrong about this. The data to prove us wrong has not been published. Draw the inference.

This started as a Fed preview and turned into a piece about execution quality, which is probably the correct evolution for a desk that spends more time reading quarterly best-execution reports than FOMC transcripts. The Bowman pause phrasing was the hook. The subject underneath it was the one that costs retail traders real money and that almost no English-language forex coverage treats with the specificity it deserves. If we rewrote this tomorrow with a different Fed governor's name attached, the second H2 would still be the one worth keeping.