Reality Index Inflation Brief · April 2026 · Methodology v0.2
The discrepancy is structural.
131 months of data going back to 2015 show the Reality Index has run hotter than the BLS CPI-U in 95% of months. The gap held at +115 bps before Covid and +140 bps after — only 25 bps apart. The April 2026 print is one of the rare months it didn't — falling mortgage rates pulled the index slightly below CPI for the first time since 2023.
Reality Index · April 2026
5.54×
$100 of 1980 goods → $554
Official CPI · April 2026
4.04×
$100 of 1980 goods → $404
RI YoY (Apr 26 vs Apr 25)
+3.57%
vs CPI +3.81%, gap −24 bps
Gap since 1980
37.1%
RI faster than CPI, cumulative
The Reality Index — a fixed-weight measure of the prices American households actually pay across ten categories of everyday spending — rose 3.57% in April 2026 compared with April 2025. The BLS Consumer Price Index for All Urban Consumers (CPI-U), unadjusted, rose 3.81% over the same period. April was the rare month where CPI ran modestly hotter than the Reality Index, driven by easing mortgage rates that moderated the cost-of-homeownership side of the index. The longer-run picture is unchanged.
This brief uses the Reality Index housing engine v0.2, which reconstructs the cost-of-homeownership component from monthly mortgage rates, the FHFA House Price Index, and Census homeownership data — replacing the carry-forward simplification used in the inaugural April brief. Headline structural-gap statistics are restated below; the longer-run divergence is sharper under v0.2, not weaker.
What 11 years of monthly data show
The Reality Index has run hotter than the BLS CPI-U in 125 of 131 months (95%) since January 2015. The mean gap is +129 basis points; the median is +132 bps.
More important than the headline averages is what happens when the data is split into two regimes: the gap held essentially constant across both calm pre-pandemic years and inflationary post-2020 years. The 2015–2019 mean gap was +115 bps. The 2020–2025 mean gap was +140 bps. Only 25 basis points apart. The structural pattern doesn't depend on the macroeconomic regime.
The Reality Index runs ~130 bps hotter than CPI in calm times and inflationary times alike. The divergence isn't a Covid story. It's a methodology story.
Reality Index vs BLS CPI-U · January 2015 – December 2025 (131 months)
The structural gap predates the pandemic.
In calm pre-pandemic inflation years (2015–2019), Reality Index ran 115 bps hotter than CPI on average. After the pandemic shock, the gap is 140 bps — only marginally wider. The post-2022 divergence isn't a new phenomenon; it's a continuation of a structural pattern that's been visible since 2015.
RI hotter than CPI
95%
125 of 131 months
Mean gap, full period
+129 bps
Median +132 bps
Pre-pandemic gap
+115 bps
2015–2019, n=60
Pandemic-era gap
+140 bps
2020–2025, n=71
Reality Index YoY
BLS CPI-U YoY (NSA)
131 months · Oct 2025 excluded (BLS data suppression)
What the long history shows: The Reality Index has run hotter than CPI in 95% of months across the full 11-year window. The 115-bps pre-pandemic gap and 140-bps pandemic-era gap are within 25 bps of each other — the structural divergence is not a post-Covid story, it's been there since 2015. In calm inflation periods (2017–2019, 2024–2025), Reality Index sits steadily 100–200 bps above CPI; during shocks (2020 oil crash, 2021 reopening surge, 2022 peak), CPI is more volatile while Reality Index follows the housing cycle. With the v0.2 monthly housing engine, the 2022 surge appears as a continuous trajectory rather than a step — the apparent vertical move in the inaugural brief was a v0.1 carry-forward artifact, now corrected.
The April 2026 print, zoomed in
The four months of 2026 published so far (January–April) sit at the right edge of the chart above. Reality Index cooled gently across the quarter, from +3.99% in January to +3.57% in April, as mortgage rates eased from a year ago. CPI climbed in the opposite direction — from +2.39% in January to +3.81% in April — and crossed above the Reality Index in April. The gap moved from +161 bps to −24 bps across four months. Both lines printed in the high 3% range by April; the index that is normally the higher one printed lower:
Reality Index vs BLS CPI-U — 2026 monthly
CPI accelerated past a moderating Reality Index.
Year-over-year change, January–April 2026. The Reality Index cooled by 42 basis points as mortgage rates eased; CPI rose by 142 basis points. The two crossed in April — only the seventh such crossing in 131 months of the backfilled history.
Reality Index YoY
BLS CPI-U YoY
Gap between the two
What drove the April print
Three buckets did most of the work in April:
Housing — the largest single weight in the index at 37.4% — rose 2.49% year over year. With v0.2 of the housing engine, the bucket combines a current-month new-buyer cost (built from Freddie Mac's PMMS rate, +1.1% tax, +0.35% insurance, +1.0% maintenance, on a home valued at the FHFA-scaled 2024 median), a 12-year vintage-MA existing-owner cost, and a renter cost (Zillow ZORI). All three responded to the falling PMMS rate (6.72% → 6.33% year over year): new-buyer cost printed −1.09%, vintage-MA existing-owner cost +3.16%, and renter cost +1.73%.[1]
Transportation rose 3.28%, contributing +0.57 pp to the headline.
Healthcare held a +5.56% year-over-year line from the most recent KFF annual employer survey baseline, contributing +0.48 pp.[2]
Utilities rose 5.39%, the hottest of the BLS-driven monthly buckets, driven by the Energy Services subindex.
Communications continued its long-running decline (−1.99% year over year), the only bucket whose price level is below where it was a decade ago. Dining out (+3.55%), groceries (+2.89%), education (+2.88%), pets (+2.72%), discretionary (+2.34%), and transportation (+3.28%) all printed in the broad +2% to +4% band that has characterized inflation since the post-2022 reacceleration.
Reality Index — April 2026 bucket breakdown
What drove the +3.57% YoY print
Each bucket's contribution to the headline year-over-year change, in percentage points. Bars sum to the headline.
Fresh monthly data (incl. housing v0.2)
Carried from 2025 annual baseline
The three largest contributors — housing, transportation, healthcare — together account for 1.98 of the headline 3.57 percentage points; the index aggregation method (ratio of weighted means rather than weighted sum of YoYs) accounts for the small residual. Healthcare and transportation are still held flat at their 2025 annual values in this edition. Housing's owner side, carried in the inaugural April brief, is now computed monthly with v0.2: fresh Freddie Mac PMMS for the new-buyer rate, vintage-MA updated for the existing-owner rate, fresh ZORI for the renter side, FHFA HPI carried within the quarter. Carry-forward is now ~26% of weight (down from ~50% in v0.1).
Why the Reality Index runs hotter than CPI
The Reality Index is built from BLS subindex data — the same source CPI uses — with three deliberate departures: fixed weights from the 2024 Consumer Expenditure Survey held constant year over year, a custom tenure-weighted housing component combining the all-in cost of homeownership with rent, and healthcare priced from the KFF Employer Health Benefits Survey rather than the BLS medical care subindex.[3]
The April 2026 print was the rare exception, not the rule. Looking back over the full 11-year window of 131 backfilled months, the Reality Index ran above CPI in 95% of months, and the cumulative gap since 1980 stands at 37.1%.
Cumulative context
Since 2000, the Reality Index has risen 2.40×. The BLS CPI has risen 1.87× over the same period. A household spending $100 a month on the Reality Index basket at the start of 2000 would now spend about $240 a month on the same basket; under CPI the figure would be roughly $187.
Since 1980, the Reality Index has risen 5.54× compared with CPI's 4.04×. The gap — 37.1% — is the cleanest single statistic for what the Reality Index is built to measure: the difference between what inflation indices report and what households experience.
Footnotes
[1] Housing methodology, v0.2. The Reality Index is published as an annual series anchored to 1980 = 100; monthly briefs extend it to monthly frequency. The housing engine v0.2 reconstructs the bucket bottom-up each month from raw inputs: (a) Freddie Mac's PMMS 30-year mortgage origination rate, averaged across the print month's weekly observations, feeds the new-buyer cost; (b) a 12-year, 0.92-decay vintage moving average of PMMS rates (refreshed each month using year-to-date PMMS) feeds the existing-owner cost; (c) the FHFA House Price Index All-Transactions for the most recent quarter scales the 2024 median home price ($420,000) to give the current home value; (d) Zillow ZORI gives the current month's renter cost (spliced onto BLS Rent of Primary Residence for pre-2015); (e) the most recent Census Housing Vacancy Survey homeownership rate weights the owner-renter blend. The owner-side cost itself blends 10% new-buyer with 90% existing-owner (the approximate annual repricing share — existing-home sales plus refinances). The reconstructed annual series reproduces the published Reality Index housing bucket within 0.16% at the 2025 endpoint and 0.03% mean absolute error across 1980–2025. Full housing methodology including formulas, parameters, and calibration is documented at housing.html.
[2] Healthcare update cadence. The healthcare bucket reflects the Kaiser Family Foundation Employer Health Benefits Survey, which is published annually each fall. Between annual releases, monthly briefs report the bucket as a carry-forward at its most recent annual level. KFF data is used in place of the BLS medical care CPI subindex because what households actually pay (premiums, deductibles, and out-of-pocket costs combined) has risen considerably faster than the BLS measure.
[3] Monthly extension methodology, v0.2. The Reality Index is published as an annual series. The monthly brief extends it to monthly frequency using four mechanisms. (1) Roughly 20% of the index by weight — dining out, communications, recreation, and pets — is reconstructed exactly from BLS monthly subindex data using a verified constant rescale to the 1980 anchor. (2) Roughly 17% of the index by weight — food at home, energy services, and education — is estimated by applying the BLS subindex's year-over-year growth at the print month to the 2025 annual baseline. This is a close approximation; the underlying Reality Index bucket may use a blend of series rather than a single subindex. (3) The full housing bucket (37.4% weight) is computed monthly with the v0.2 sub-component engine described in footnote [1]. (4) Transportation and healthcare (together 25.9% of weight) are still carried forward at their 2025 annual values for this edition. All comparisons use not-seasonally-adjusted data (April 2026 vs. April 2025) to neutralize seasonality without the obscuring effect of seasonal adjustment. Fresh data accounts for ~74% of index weight in this print month; the balance carries from 2025 — down from ~50% in the v0.1 monthly engine.
[4] The 11-year backfill. The chart at the top of this brief reconstructs the Reality Index at monthly frequency from January 2015 through December 2025 — 131 months — using housing engine v0.2 for every month. Each month is anchored to its own prior-year annual baseline (e.g., 2023 print months anchor to the 2022 annual). The 12 months of 2015 use the BLS Rent of Primary Residence (CUUR0000SEHA) as the renter signal since Zillow ZORI begins January 2015 and a prior-year ZORI comparable is needed; 2016 onward uses ZORI directly. October 2025 is excluded because BLS published "−" (suppressed) for one of the underlying subindexes that month. The reconstruction uses current-vintage BLS data, meaning each month's number reflects today's BLS file rather than what was reported at the time of original release — standard practice for retrospective series. Compared with the v0.1 inaugural backfill, v0.2 catches the 2021–22 housing surge and the 2023 housing moderation that v0.1's annual carry-forward smoothed away. The 11-year structural-gap headline statistics shift slightly: RI hotter than CPI in 95% of months (up from 84% in v0.1), mean gap +129 bps (up from +122), pandemic-era gap +140 bps (up from +128). The pre-pandemic gap is unchanged at +115 bps to two decimal places — the structural divergence is methodology-agnostic.
Sources: BLS CPI-U (NSA), Zillow ZORI, FHFA HPI, KFF Employer Health Benefits Survey, U.S. Census Housing Vacancy Survey. Full methodology: methodology_paper.html.
Future monthly editions will appear here. The next edition will follow the May 2026 BLS CPI release.