On the Eve of CPI Release: The Widening Gap Between Official Inflation Data and Consumer Reality
0xBroomberg
U.S. May CPI ran at 4.2% year-on-year, yet the Michigan consumer-sentiment index hit its lowest since 1978 — official data says 'no crisis,' but households feel squeezed, and that gap is shaking confidence in single-number inflation readings.
Why do the data and the feelings disagree?
May CPI came in at 4.2% year-on-year; PCE — the Fed's preferred inflation gauge — at 3.4%. The official picture: "concerns, but no crisis."
Yet the University of Michigan consumer-sentiment index hit its all-time low since records began in 1978 in May; June's reading was the second-lowest ever.
This means → inside the same economy, the statistical thermometer and the household thermometer have diverged systematically. The "moderate inflation" investors see and the price pressure consumers live with may be two different realities.
What does the "average" actually hide?
The Bureau of Labor Statistics tracks roughly 100,000 goods and services each month, weights them by spending surveys, and produces a CPI meant to represent the "typical consumer." It currently maintains just three market baskets.
Labor economist Kathryn Anne Edwards argued in a Bloomberg column that this framework compresses wildly different consumer groups into a single mean.
In plain terms = one household earns $30k a year, another earns $300k. Their lived experience of prices is vastly different — but CPI hands you one middle number that reflects neither.
How big is the gap?
The BLS's own research shows: between 2006 and 2023, the lowest-income quintile faced annual inflation roughly 0.28 percentage points higher than the highest-income quintile.
That sounds small — but compounded over nearly two decades, it adds up to 7.7 percentage points.
This means → low-income households absorbed far more cumulative price pressure than wealthy ones, and that wedge is almost invisible in headline CPI. Policymakers watching the aggregate number are reading the economy through a blurred lens.
Is this hard to fix?
Edwards's core point: the heaviest work — collecting 100,000 price readings each month — is already done.
From the same raw data, re-weighting by household type, income level, renter vs. owner status, and age would generate a much richer set of sub-indices. In plain terms = the data already exist; the Bureau just slices the cake too coarsely. A finer knife is all it takes.
She recommends expanding the current three baskets by at least tenfold and publishing monthly data for each typical household type.
Would better data be enough?
Edwards is explicit: improving the measurement cannot fix the economy's underlying problems.
The U.S. economy faces multiple structural pressures: slowing hiring, stalling wage growth, persistently elevated prices, rising credit-card debt, high rates suppressing the housing market, and AI's looming impact on jobs.
This reflects a deeper reality — consumer confidence at historic lows is not a reaction to any single number but a composite sensation of overlapping pressures. A single CPI print cannot capture that compound anxiety.
June CPI is days away — how should investors read the number?
The core question: how well can one aggregate indicator capture the real divergence in inflation pressure across this economic cycle?
This means → a "moderate" June CPI print would not mean consumer-side risk has faded — the pressure underneath, stratified by income, may simply be averaged away.
For reading the Fed's policy path, the divergence itself in consumer behavior is the key variable — not a statistically smoothed headline number.
Content is for reference only, not financial advice.