EURUSD: The Same Bottom as March 2025. 1.13706 Kills ItEUR/USDOANDA:EURUSDIvanLabrieA weekly buy signal fired on this pair off Friday's close. The levels are further down. First, why this is not the trade the tape looks like it should be handing you. FX is mainly a relative monetary policy story, and what drives the trend is the direction and the rate of change of that gap rather than its level. Other things move a pair, but rarely for long and rarely this far. So the question is not which central bank is hawkish, because both of these are. It is which one still has repricing left to do. The corollary matters more than the rule. When everyone is easing at the same pace, or hiking at the same pace, there is no gap to trade and FX turns into a choppy, untradeable mess. You do not get paid for having a view on it. You get paid when the gap starts moving again. That is the whole bet here. Everyone can see the hawkish Fed. It held at 3.50 to 3.75% on a 9-3 vote with all three dissents FOR a hike, the most policymakers pushing for an increase since 2016, and it dropped forward guidance entirely. On the face of it, dollar-positive. It is also finished being priced. September sits at 57.4% for a hike, October 66.7%, December 80.5%, and the probability of an ease at ANY meeting this year is exactly zero. The two-hikes-by-September bucket went from 20.2% to nothing in a single week. There is no dovish Fed left to remove and no hawkish Fed left to add. A path that is fully priced stops paying the currency carrying it. What has not been priced sits on the other side of the pair. Eurozone July CPI came in at 2.9%, up from 2.8%, and it went up because energy costs lifted it, with core and services both accelerating. A month ago a September ECB hike was not a live question. It is now the focus. One cost-push shock, three central banks, all at different stages of answering it. The Fed has taken its three dissenting votes already. The Bank of England held and split 6-3 hawkish, three votes to hike and none to cut, with an explicit path to 4.5% if oil holds. The ECB has not repriced, and it is the one sitting closest to the energy input, because Europe imports the barrel it is being taxed by. The marginal hawkish surprise has moved from the dollar side of this pair to the euro side. That is the whole idea. There is a second tell, and it is the one that keeps me interested. The US 30-year is at a 19-year high, and it got there in the same week September hike odds were PARED BACK rather than added to. Yields rising while the expected policy path softens is a curve steepening on fading credibility, not a hawkish repricing. A currency does not usually get paid for that. Same distinction as a stock going up on the multiple instead of on the earnings, and it tends to resolve the same way. The American data underneath was a stagflation print. Q2 GDP 1.5% against 1.8% expected. Jobless claims at 197,000, so labor is not the crack and the softness is growth. June core PCE 0.1% on the month against 0.2% expected, and the timing there matters more than the number, because June predates the crude spike entirely. A soft June core is what the oil-lag read predicts, not evidence against it. Fully priced hiking path, decelerating growth, and the euro side picking up an inflation problem of its own in the same window. That is the setup. Now the levels, which are the part you can hold me to. The signal comes off the week of July 27, which closed at 1.15306, up 1.41%. Partial exits sit at 1.15490, 1.17068, 1.18646, 1.20224 and 1.21802. The last of those is 5.63% above the close. The level that kills it is 1.13706, 1.41% under the close, and if that goes the read is wrong and there is nothing to argue about. One detail off that bar worth having. The week's high was 1.15476. The first target is 1.15490. It stopped 1.4 pips short and closed 18.4 pips under it. Whatever you make of the macro, the bar itself ran straight into the first level and parked under it. This usually translates into higher probability of accelerating towards target #2 or higher compared to signals where the pace is slower initially. On the chart, the vertical is the signal bar. The arrows are swing turning points, not entries. Now the part I actually find interesting. Technically this is the same kind of weekly bottom the euro made in March 2025. Same shape, same position after a long grind lower, same look on the bar that turned it. That signal ran for the best part of a year before this pair topped out in February. I am not going to pretend the two setups are identical. But if you have been trading this pair for a while, that bar in early March last year is the one this reminds me of, and the last one worked. There is one more thing about that March 2025 bar. The week it printed was the week Germany tore up its own fiscal rules, a 500 billion euro infrastructure fund plus a debt-brake exemption for defence. Ten-year Bund yields rose 43 basis points that week to 2.84%, the largest weekly jump since reunification in 1990. The euro went from 1.0178 in January to 1.0955 by 18 March and kept going. Nothing the Fed did was involved. The euro was repriced by European rates. I am not claiming a second fiscal bazooka, because there is not one. The rhyme is not the spending, it is the channel. European rates moved the euro then without any help from Washington, and European rates are what would move it now, with the ECB doing the job the Bundestag did last time. Which is why I care about the shape of the bar more than I normally would. Two ways this goes. The one I lean to: the ECB spends August walking toward September the way the Fed spent July walking toward its dissents, and the pair grinds into the target range on differential compression that needs no dovish word out of Washington. Nothing has to break for this one. Europe only has to reprice what has already happened to its own CPI. The other one is live and I am not going to talk around it. The energy shock reverses, European inflation rolls straight back over because the whole move up came off the energy line, and the September ECB question closes before it properly opens. Then the differential goes back to being about US growth alone, and 1.13706 is where I find out. One note on what this post is. It is a signal I am publishing because it belongs on the record, not a book I am talking. I am not in this one. If it works the timestamp is the argument, and if it does not, the invalidation is sitting in the same post, published the same minute. Macro and Time@Mode, same as always. Everything I publish is timestamped. Cheers, Ivan Labrie.