The Exness Pro account advertises a 0.1 pip average spread on EUR/USD. IC Markets Raw quotes the same neighbourhood, and Pepperstone Razor and Tickmill Pro sit inside a few hundredths of a pip of that print. Those are the receipts posted on the marketing pages, and they are the numbers a retail desk models when it sizes a yen trade. This piece works the actual arithmetic of what happens to that 0.1 pip when the Bank of Japan delivers a hawkish surprise and the yen surges. The receipt is 0.1. Everything after explains how far it moves.
What the Numbers Actually Say
At 11:47 GMT on the morning of a BoJ decision — a real one, the March 2024 exit from negative rates, or any of the yield-curve-control adjustments that came before it — a retail dashboard shows one number. The number is the advertised average. It is not the number that clears the trade at the moment the print hits.
Start with the receipts as they are marketed. Exness Pro: 0.1 pip average on EUR/USD. FBS Pro: 0.0 pip average, plus a commission. HF Markets Pro: 0.0 average, plus a commission. FXTM Pro: 0.1. AvaTrade does not run a raw model — 0.9 pip standard, wider by design because the spread is the revenue. IC Markets Raw and Pepperstone Razor and Tickmill Pro cluster at the same 0.0 to 0.1 print, and they charge a commission — the commission-plus-raw architecture the industry converged on after the 2001 electronic-trading transition compressed markup spreads from the five-to-ten pip range of the manual dealing desks down to fractions of a pip.
The USD/JPY quote is not the EUR/USD quote. USD/JPY trades wider on average than EUR/USD, even at the same broker, on the same account tier, in the same session. Volume ratios explain most of the gap — EUR/USD is the deepest book in the market, USD/JPY the second, and the ratio between the two shows up in the tick data as roughly a 30 to 60 per cent spread premium for the yen cross. That is the pre-event baseline. A 0.1 pip EUR/USD print translates, in most retail books, to something between 0.15 and 0.25 pips on USD/JPY when the session is calm.
Calm is the qualifier that does all the work here. The 0.1 average is not a floor. It is an average across the quoted window — 24 hours in most cases, weighted by tick count. The tail of that distribution — the top one per cent of ticks — sits nowhere near 0.1. This is not fine print. It is the actual shape of the number.
The BoJ helpline for market participants is not something a retail trader calls. The BoJ meeting statement drops on the wire and the price adjusts before the PDF finishes downloading.
What Nobody Mentions
The advertised average is a marketing artefact. It is truthful and it is misleading in the same breath. What the marketing does not surface is the spread distribution — the shape of the histogram behind the average.
Consider the mechanics. A Pro account or a Raw account routes to a liquidity aggregator. The aggregator pulls quotes from a panel of tier-one bank dealers and non-bank market makers. In quiet conditions, the best bid and best offer from the panel sit inside a fraction of a pip — because dealer risk is low, inventory is easy to hedge, and the competition among market makers compresses the top of book. In an event window, that compression evaporates. Dealers widen. Some pull out of the panel entirely for thirty to ninety seconds. The aggregator still returns a quote, but the quote is the surviving inside price, not the calm-window inside price.
Historical reconstruction is useful here. During the 2001-2004 transition from voice broking to electronic ECN dominance, EBS and Reuters Matching published aggregate spread data for interbank USD/JPY. The 2001 mean was around 3 to 4 pips at the top of book. By 2004 it had compressed to under 2. By 2010 the interbank inside sat under a pip in normal hours. Retail markup, layered on top, moved in parallel — the five-to-ten pip retail spreads of the late 1990s dealing-desk era did not survive the pass-through pressure of ECN competition. The commission-plus-raw model at brokers like IC Markets Raw and Pepperstone Razor and Tickmill Pro is the direct descendant of that transition. It exists because the interbank number is small enough that a retail markup is embarrassing to advertise.
But the tail of the distribution never compressed the way the mean did. Interbank USD/JPY during the 2013 QQE announcement, during the 2016 negative-rate shock, during the October 2022 intervention, and during the March 2024 exit from negative rates all showed the same pattern in the public tick data: inside spreads that were sub-pip in the quiet minute before the wire ran, and multi-pip in the sixty seconds after. Widening ratios of 20x to 60x over baseline are documented in BIS working papers on FX microstructure during announcement windows.
A retail Pro or Raw account inherits that widening. The account tier does not insulate the trader from event microstructure — it inherits it in unfiltered form. That is the trade-off of raw pricing: you get the tight quote in the quiet window and you get the wide quote in the event window, without a smoothing markup to average them out.
The BoJ press conference runs about forty minutes. The dealer panel does not fully re-tighten until the governor stops speaking.
The Real Cost
Now do the arithmetic.
Take a one standard lot USD/JPY trade — 100,000 units of USD. The pip value on USD/JPY at a spot rate near 150.00 is approximately $6.67 per pip per standard lot. Round to $7 for a conservative working figure.
Baseline case, calm session, Pro account on EUR/USD is 0.1, USD/JPY runs about 0.2. Round-turn cost is 0.4 pips — 0.2 in, 0.2 out. In dollar terms, 0.4 × $7 = $2.80 per standard lot. Add a commission of roughly $7 round-turn on a Raw or Pro account, and the total round-turn cost is under $10 per standard lot in normal conditions. That is the number the retail desk models when it sizes a yen trade.
Now the hawkish-BoJ scenario. Assume the BoJ delivers a surprise — a policy rate move up, a YCC band adjustment, or a signalled shift in the balance sheet path. The wire hits. Yen surges. The USD/JPY quote goes from a calm 0.2 pip inside to something between 3 and 10 pips inside for the next thirty to ninety seconds.
At 5 pips of event-window spread, the round-turn cost on entry-and-exit inside that window is 5 + 5 = 10 pips. 10 × $7 = $70 per standard lot. Add commission of $7. Total $77 per standard lot. That is 7.7x the calm-window cost.
At 10 pips — the upper end of what the tick record shows for USD/JPY in the peak volatility minute of an event print — the round-turn is 20 pips. 20 × $7 = $140. Plus commission, roughly $147 per standard lot. That is nearly 15x the calm-window figure.
Scale the trade. A retail account running ten standard lots — not uncommon for a modestly leveraged carry desk running a yen basket — pays between $700 and $1,470 in spread on a round-turn taken inside the event window. That is not the P&L of the trade. That is the toll on entering and exiting the position at the wrong second.
There is a second cost that does not show up in the spread column. A stop loss placed inside a Pro or Raw quote stream during an event window can execute at the wide inside — meaning the fill price is not the level the retail dashboard displayed pre-event. Slippage on stops during the March 2024 BoJ decision was documented in retail broker execution reports at levels between 8 and 25 pips on USD/JPY. That slippage stacks on top of the spread. A stop placed 20 pips from entry, hit during the event window, can clear at 45 pips from entry. In dollar terms, on a ten-lot position, that is an additional $1,750 versus the modelled loss.
The 0.1 pip receipt on the marketing page is honest — for the window it describes. The window it describes is not the window in which a BoJ decision is priced.
If You Only Remember One Thing
The advertised spread is an average of a distribution, not a floor. In the tail — the announcement minute, the intervention minute, the surprise print — the spread widens by an order of magnitude, and the retail account tier does not insulate you from that.
The math is not hidden. It is public in BIS microstructure papers, in broker execution reports, in the tick data any Pro account can pull from its own platform. It just does not appear on the marketing page. Whether the average print or the tail print is the honest number to model a yen trade around depends on when you plan to trade. If you plan to trade the BoJ, model the tail. If you plan to trade the tail, ask a question the retail industry has not answered publicly yet: what is the actual per-broker slippage distribution on USD/JPY inside the event window, per liquidity provider, across the last five BoJ meetings? The aggregated number exists somewhere in the aggregators' logs. If you have seen it published in a form a retail trader can act on, write.
FAQ
How much does a round-turn USD/JPY trade actually cost at a Pro or Raw account in normal conditions?
On a one standard lot USD/JPY trade at a spot rate near 150.00, calm-window spread is roughly 0.2 pips each way — a 0.4 pip round-turn. At a pip value of about $6.67 per standard lot, that is under $3 in spread cost. Add a Pro or Raw commission of roughly $7 round-turn per lot, and the all-in cost sits under $10. That is the modelling number. It is only correct outside event windows.
Why is the USD/JPY spread wider than the EUR/USD spread on the same account?
Depth of book. EUR/USD is the deepest quote in the market, USD/JPY the second. The tick-data premium for USD/JPY over EUR/USD in retail Pro-tier books runs 30 to 60 per cent in calm conditions. When brokers advertise a 0.1 pip average on EUR/USD, the corresponding USD/JPY print is closer to 0.15 to 0.25 pips. That is the pre-event baseline before microstructure widens further on any announcement.
What happens to the spread during a BoJ decision or hawkish surprise?
The dealer panel behind the liquidity aggregator widens quotes for thirty to ninety seconds after the wire. Some market makers exit the panel briefly. The surviving inside spread on USD/JPY has historically moved from sub-pip to somewhere between 3 and 10 pips during announcement windows — a widening ratio between 20x and 60x over baseline, documented in BIS working papers on FX microstructure and in retail broker execution reports.
Are Raw and Pro accounts safer from event-window spread widening than standard accounts?
No, structurally the opposite. Raw and Pro accounts pass through the aggregator's inside quote without a smoothing markup. That means you get the tight quote in quiet conditions and the wide quote in event conditions, unfiltered. A standard account like AvaTrade's 0.9 pip fixed-ish print averages the two windows into a broader baseline. The Raw model is cheaper on average and more punishing in the tail.
Does the commission stay flat during event windows or does it also widen?
The commission stays flat. Commission at commission-plus-raw brokers — IC Markets Raw, Pepperstone Razor, Tickmill Pro — is a fixed per-lot fee, typically around $7 round-turn on a standard lot. It does not adjust with volatility. The variable component of the cost is entirely in the spread. That is why the round-turn cost of a trade taken in an event window can multiply while the commission line stays unchanged.
How much can stop-loss slippage add to the modelled cost on a hawkish BoJ print?
On the March 2024 BoJ decision, retail execution reports showed USD/JPY stop slippage in the 8 to 25 pip range during the peak announcement minute. On a ten standard lot position, that is between $560 and $1,750 of unmodelled fill cost, on top of whatever the spread was at the moment of execution. Stop-loss levels set pre-event do not necessarily clear at the displayed level once the aggregator's inside widens.
Did retail spreads always look like this, or is the compression recent?
The compression is recent. Pre-2001, retail forex ran on dealing-desk markups of 5 to 10 pips on major pairs, because voice broking and manual quoting kept interbank spreads wide. The 2001-2004 transition to electronic trading — EBS and Reuters Matching taking over interbank USD/JPY — compressed the top of book, and ECN pass-through pressure forced retail markups down in parallel. The commission-plus-raw account tier that dominates today is the residue of that transition.
Which is more important to model — the average spread or the event spread — for a yen trade?
It depends on when you intend to trade. For carry positions held across weeks, the average spread dominates the cost basis and the marketed 0.1 pip figure is a reasonable proxy. For any trade taken through a BoJ meeting, an intervention window, or a scheduled policy print, the event spread and the associated stop slippage are the relevant numbers. Using the marketed average to model an event trade understates the cost of the round-turn by roughly an order of magnitude.