Headline CPI vs Headline PCE

Headline CPI (CPIAUCSL on FRED) is the Consumer Price Index for All Urban Consumers, published by the BLS on the second Tuesday of each month. It uses a Laspeyres-style fixed-basket framework with weights updated every two years from the Consumer Expenditure Survey. Headline PCE (PCEPI on FRED) is constructed by the BEA from Personal Consumption Expenditures inside the National Income and Product Accounts, releases on the last business day of the month, and uses a Fisher-ideal chain index that updates weights monthly to reflect substitution.

The two indices are built on different source data. CPI surveys what households tell BLS they spent. PCE pulls from retail-sales receipts, business surveys, and administered-pricing schedules, which is why employer-paid health insurance and Medicare reimbursements show up in PCE but not CPI. The Fed's 2012 statement of longer-run goals defines the 2% target in terms of headline PCE precisely because PCE covers a broader expenditure universe and adjusts faster to consumer behavior. As of the BEA release for March 2026, headline PCE printed 3.5% year-over-year against headline CPI at 3.3%, with the BEA reporting the monthly gain at 0.7% versus the BLS monthly print of 0.9% on a seasonally-adjusted basis.

Shelter is the dominant mechanical wedge between the two series. As of January 2025 the BLS reported shelter at 35.4% of headline CPI; the BEA's most recent rebasing puts housing at roughly 15.6% of headline PCE. That weight gap means the same Owners' Equivalent Rent print contributes more than twice as much to CPI as to PCE on impact. When OER ran above 7% from mid-2022 through mid-2024, this single category pushed headline CPI well above headline PCE for 30 consecutive months.

Healthcare runs the other way. PCE incorporates third-party payments through Medicare, Medicaid, and employer-sponsored insurance, and total healthcare reaches roughly 17% of headline PCE. CPI healthcare reflects only out-of-pocket consumer spending and weights at about 8%. Medicare reimbursement schedules are set administratively and lag market rates, which is why physician-services inflation can print 70 to 100 basis points apart in the two series in any given quarter. The August 2024 BLS-vs-BEA divergence on physician services was nearly 90 basis points, almost entirely attributable to Medicare's 2024 fee-schedule update flowing through PCE but not CPI.

The March 2026 release was the cleanest example in two decades of headline PCE running above headline CPI for non-methodological reasons. CPI printed 3.3% year-over-year; PCE printed 3.5%. Both gains were dominated by energy: gasoline rose 21.2% in March alone, contributing roughly 75% of the monthly CPI increase, after the Iran conflict pushed Brent crude above $103 and WTI to $95.85 by April 23. The Trump administration's March 11, 2026 release of 172 million barrels from the Strategic Petroleum Reserve dampened the rally but did not reverse it.

PCE topped CPI because the energy shock hit a basket where shelter (a CPI-heavy category) was already disinflating. Shelter rose just 0.3% monthly and 3.0% year-over-year in March 2026, the lowest since August 2021, which pulled CPI down faster than the broader PCE basket. The Cleveland Fed inflation nowcast and the Dallas Fed Trimmed-Mean PCE both flagged the inversion as energy-driven and likely to reverse if WTI returns below $80, but for the May 7, 2026 FOMC meeting it leaves the Fed staring at a 3.5% PCE print against a 2% target.

From 1979 through 1981, headline CPI peaked at 14.6% in March 1980 while headline PCE peaked at 11.6% in the same window, a gap of nearly 300 basis points driven mostly by CPI's pre-1983 treatment of mortgage-interest costs. The 1983 BLS methodology change that replaced mortgage interest with rental equivalence closed roughly 150 basis points of the structural gap.

The 2008 oil shock produced a reverse-direction divergence: headline CPI peaked at 5.6% in July 2008 while headline PCE peaked at 4.1%, then both collapsed into deflation territory by mid-2009 with CPI hitting -2.1%. The 2021-2022 episode put CPI at 9.1% in June 2022 against PCE at 7.0% the same month, the widest positive CPI-PCE gap in the modern series. The 2022 oil round-trip pulled CPI back to 3.0% by June 2023 while PCE took two additional months to follow because of healthcare-administered-pricing lags. The March 2026 inversion, with PCE above CPI under an energy-supply shock, has only one close historical analog: the brief mid-2015 episode when collapsing oil and a strong dollar pulled headline CPI to -0.2% while PCE held at +0.3%.

Headline CPI prints first, around the 10th of each month at 8:30am ET, and is the public anchor. Headline PCE follows roughly two weeks later. Macro desks build a headline-PCE nowcast from CPI plus PPI components that map into PCE healthcare and financial services, then trade the residual. The Cleveland Fed's published Inflation Nowcasting series is the institutional benchmark for this workflow, and as of late April 2026 it was running consistent with the March BEA print within roughly 5 basis points.

The market reaction asymmetry is informative. The April 10, 2026 CPI release at 3.3% headline produced a 7-basis-point move in 2-year Treasuries despite the historic 0.9% monthly gain, because energy-driven prints are partially looked through. The April 30, 2026 PCE release at 3.5% produced a 12-basis-point repricing of the December 2026 SOFR contract, because PCE feeds directly into the SEP and the dot plot. TIPS breakevens accrue on headline CPI specifically; using PCE expectations to evaluate TIPS prices is a structural error that produces 30 to 40 basis points of model bias over a typical cycle.

The March 2026 SEP projected headline PCE at 2.7% for year-end 2026, compared to long-run goal 2.0%, with the energy shock cited explicitly in the meeting minutes as the reason for the upward revision from December's 2.4%. Both the headline and the core PCE projection sit above target through 2027 in the latest dot plot, which is why the Fed has held the policy rate at 3.75% since the December 2024 cut despite Sahm Rule signaling.

For portfolio positioning, the operating rule is that headline CPI is what households feel and Social Security cost-of-living adjustments reference, while headline PCE is what determines policy. The 2026 COLA was set at 2.5% based on Q3 2025 CPI-W; the 2027 COLA window opens July 2026 and the Iran-driven energy bump puts it on track for 2.7-3.0% pending oil-price persistence. Investors who confuse the two series on policy timing will systematically misread the Fed reaction function. Reading both together, with the Cleveland Fed nowcast as the institutional cross-check, is the workflow that survives the March 2026 inversion intact.