US core PCE price index, May 2026, month-over-month % change

Core PCE is the Fed's preferred inflation gauge and is comparatively smooth. Recent prints show a clean three-month deceleration toward the trailing mean (~0.24%). The forecast holds near that level with a tight CI given low realized volatility.

Point: trailing-24 mean (0.243%) and the decelerating last print (0.24%) coincide -> 0.25% (rounded). 80% CI from realized vol: half-width = 1.28*0.096 = 0.123pp; widened to 0.15pp for rounding and tail risk. Interval = 0.25 +/- 0.15 = [0.10%, 0.40%].

What lands it outside [0.10%, 0.40%]: a re-acceleration in core services/supercore (e.g. a portfolio-management or healthcare-services jump) or a tariff-driven core-goods pass-through pushing it above 0.40%; or an unusually soft services month / negative goods print dropping it to 0.10% or below. The Fed watches this number most closely of the seven.

The control holds near the recent core PCE trend and does not ingest CPI components that arrive before the BEA release.

The CPI bridge raises the center modestly because core CPI source data is firm enough to keep BEA core PCE above the no-pack persistence estimate.

The resolver is the BEA first print for Personal Income and Outlays, May 2026: PCE price index excluding food and energy, seasonally adjusted monthly percent change. This is a first-print forecast, so later BEA revisions should not alter the resolved value.

Base-rate/reference-class: recent core PCE monthly prints cluster between about 0.2 and 0.4 percent, with the FRED-implied latest four monthly changes at 0.239, 0.295, 0.396, and 0.427 percent. The April first print rounded to 0.2 percent despite firm headline PCE, so a May forecast should lean below the raw recent average but above the softer core CPI alone.

Counter-consideration: May core CPI at 0.2 percent and declines in new vehicles, household furnishings, and motor vehicle insurance point to a low core PCE print. Against that, PPI services and portfolio-management prices were firm, and some PPI inputs feed more directly into PCE than CPI components do.

Recent FRED-implied core PCE mean for Jan-Apr is (0.427 + 0.396 + 0.295 + 0.239) / 4 = 0.339 percent. I downweight that toward May core CPI of 0.2 because CPI was softer, while adding some PPI-services pressure: 0.55*0.24 latest core PCE momentum + 0.30*0.20 core CPI + 0.15*0.52 PPI-informed pressure = 0.270 percent. An 80 percent interval of 0.12 to 0.42 covers normal nowcast error and component-mapping uncertainty.