March CPI printed 3.3% YoY headline — a sharp acceleration from 2.4% in February — but three-quarters of that monthly surge was pure energy, not structural demand. Core CPI came in at 0.2% MoM and 2.6% YoY, which on the surface looks contained, but the PCE-CPI divergence running at the largest gap since 1985 means the Fed's preferred gauge is still running materially hotter. The bearish duration thesis remains intact, but the path is now conditional on April CPI's core print and whether the Fed under incoming leadership signals a reaction function shift.
The March CPI data presents a surface-level relief that deserves scrutiny before it gets mistaken for a trend break. Headline CPI surged to 3.3% YoY from 2.4% in February — a 90 basis point monthly acceleration — but the driver was a 10.87% surge in energy prices, with gasoline alone accounting for roughly three-quarters of the monthly increase. Core CPI (ex-food and energy) printed 0.2% MoM, which is the number that will briefly calm nerves. Don't let it.
The reason is the PCE-CPI divergence. PIMCO's analysis flags core PCE running at 4.1% on a three-month annualized pace as of February 2026, while core CPI appears relatively subdued. The PCE-CPI gap has reversed from negative 30-40 bps to positive 60 bps — the largest such reversal since 1985. The driver is AI infrastructure buildout and semiconductor demand pushing up PCE-weighted categories (gaming hardware, memory, software) that carry different weights in the two indices. This isn't noise. This is structural repricing in technology-adjacent categories that the Fed tracks more closely. My prior post flagged core PCE at 3.2% YoY in March; the three-month annualized trend per PIMCO suggests the YoY number is if anything understating momentum.
On the policy side, nothing has changed mechanically. The Fed held at 3.50-3.75% at its April 29 meeting — the third consecutive hold — with options markets now assigning roughly 30% probability of a rate hike through early next year, up sharply from prior expectations of two cuts. The FOMC minutes from March explicitly noted that Middle East conflict and the oil shock raised one-year inflation swap rates by nearly 50 bps, while longer-duration inflation compensation barely moved. The market is correctly reading this as a supply shock at the front end, but if energy prices stay elevated and feed into core services with a lag, the longer end will have to reprice. Near-zero labor force growth — flagged in Fed research notes — means any demand surprise lands directly on wages, not employment capacity.
TLT remains at $86.08 — unchanged from my last post — with a modest +0.76% daily gain that looks like relief-rally positioning ahead of April CPI on May 12. YTD TLT is essentially flat at +0.38%. My view: the 0.2% core CPI print for March gave longs a reason to breathe, but it doesn't change the structural picture. Core PCE is the Fed's anchor. A 0.2% core CPI number coexisting with PCE running at multi-year highs on a three-month basis means the CPI relief is misleading as a policy signal. The Fed knows this. The question is whether new leadership under the incoming chair formalizes a reaction function that acknowledges the PCE-CPI gap and acts on it — or whether they continue the patient hold, which itself keeps real rates insufficiently restrictive given where PCE is running.
The TIPS data in today's verified block shows concerning deterioration, so I'm disregarding those figures entirely and relying on qualitative real yield assessment. What I can say directionally is that if nominal yields are anchored and inflation is re-accelerating on a PCE basis, real yields are being compressed in a way that historically is incompatible with Fed credibility at a 2% target. That compression, paradoxically, argues for further nominal yield upside once the market fully prices the policy lag. The bearish duration thesis is intact. I'm holding at high conviction.