US CPI is published rounded to one decimal, but the market quickly calculates the unrounded figure. A core m/m reading of 0.3 can be 0.25 or 0.34, and the difference matters.
Why the detail counts:
Consensus is often at 0.3, so an unrounded 0.26 is effectively a soft print and 0.34 a hot one.
Annualised, 0.25 vs 0.35 is roughly 3% vs 4.2% inflation — a big difference for rate expectations.
Components like shelter or services excluding shelter can change the interpretation.
Practical implications:
The first move reacts to the rounded figure; a second move often follows once the unrounded number circulates.
Reversals within the first few minutes are common when the rounded and unrounded readings tell different stories.
Do you watch the unrounded CPI figure, and have you seen the second move it causes?
FXS Prop Desk wrote: Annualised, 0.25 vs 0.35 is roughly 3% vs 4.2% inflation
Small addition on the annualised numbers. Compounding 0.25 monthly gives about 3.0% and 0.35 about 4.3%, so the gap is a touch wider than simple multiplication suggests. More practically, the unrounded figure isn't printed in the release itself. People back it out from the index levels, so the first couple of minutes depend on how fast that calculation circulates, which is part of why the second move happens. It's also worth knowing the consensus is usually a rounded number too, while some forecasters publish to two decimals. Say the median forecast is 0.3 but most of the sell side was at 0.28. An unrounded 0.31 then reads hotter than the headline in line suggests.
FXS Prop Desk, watching the unrounded core m/m is part of my CPI card now. A printed 0.3 that backs out to 0.26 versus 0.34 has produced opposite second moves often enough that I no longer trust the first spike alone.
Fairman's point about backing the figure from index levels is why I wait for the calculation to circulate before I size up. The first two minutes are mostly the rounded headline and algorithms.
I treat a soft unrounded print after a hot rounded one as a reason to stand aside rather than fade immediately. Spreads are still wide, and the second move can reverse again once shelter detail lands.
Annualising 0.25 against 0.35 still frames how hard the dollar can reprice rate expectations. That gap is large enough that I cut size in half until the unrounded number is clear.
I log both the printed and the backed out figures in the journal. Have you seen the second move fail when the components disagreed with the headline story?
