Inflation Metrics Broke When We Started Measuring the Wrong Basket

The CPI Measures a World That Doesn't Exist Anymore

The Consumer Price Index tells us inflation is hovering around 3%, which would be reassuring if anyone actually lived in the economy the CPI measures. The basket of goods hasn't kept pace with how we actually spend money in 2024. Housing costs get smoothed into "owner's equivalent rent" that bears no relationship to what people pay for mortgages at 7% rates. Healthcare premiums jump 15% annually but get weighted like it's still 1990.

Meanwhile, the things eating our budgets—childcare, education, software subscriptions—either don't show up properly or get "hedonically adjusted" into irrelevance. The BLS decided your $200/month SaaS stack doesn't count as inflation because the software has "more features" than last year. By that logic, a $50,000 car with a backup camera isn't more expensive than a 1995 Civic.

Tech Subscriptions Became the New Rent

Somewhere between 2015 and now, software companies figured out they could raise prices 20-40% annually as long as they called it something else. Not inflation—just "pricing optimization" or "plan restructuring" or my personal favorite, "aligning value with usage."

The mechanics are brilliant in their simplicity. Adobe didn't just raise prices—they eliminated perpetual licenses and made the old software incompatible with new file formats. Salesforce doesn't increase costs—they just move features you were using into higher tiers. Slack, Zoom, Notion, Figma: the pattern repeats across every category. Switching costs are high enough that complaining is cheaper than migrating.

This shows up nowhere in inflation metrics because the BLS treats software like a commodity with perfect substitutes. In theory, if Slack gets expensive, you switch to Microsoft Teams. In practice, you've got three years of integrations, institutional knowledge, and workflow automation that makes switching cost more than just paying the increase.

Housing Inflation Is Worse Than the Numbers Suggest

The CPI's treatment of housing might be its biggest fiction. Owner's equivalent rent—the amount homeowners would theoretically pay to rent their own homes—smooths out the actual cost of housing into something statistically convenient and practically useless.

Mortgage rates tripled from 3% to 7% in eighteen months. For anyone buying a home, monthly payments jumped 60-80% even as nominal prices stayed flat. The CPI barely noticed because existing homeowners with locked-in rates didn't see their costs change. The measurement optimizes for continuity over accuracy.

Rent inflation shows up better but still lags reality by 12-18 months due to how the BLS samples. By the time rent increases hit the index, landlords are already implementing the next round. The metric becomes a trailing indicator of a crisis everyone living through it already knows about.

Services Inflation Reveals the Real Story

Goods inflation cooled because supply chains normalized and we stopped panic-buying Pelotons. Services inflation—the cost of things humans actually do—never stopped accelerating. Haircuts, restaurant meals, car repairs, childcare: anything involving labor keeps climbing at 5-7% annually.

This matters more than the headline number suggests because services are where we can't substitute down. When TVs got expensive, people waited or bought smaller ones. When childcare costs jump 30%, parents don't just hire cheaper babysitters—the market doesn't work that way. You pay or you quit your job.

The composition shift makes aggregate inflation numbers increasingly meaningless. A basket that's 60% goods and 40% services averaging 3% inflation hides that services are running at 6% while goods are flat. The people spending 70% of their income on services—which is most people under 40—are experiencing an entirely different economy than the index suggests.

Shrinkflation Became the Quiet Part We Say Out Loud

Companies discovered they could hide price increases by making products smaller, and the BLS has no systematic way to track it. The classic example is cereal boxes shrinking from 19oz to 16oz while prices stay constant—a 16% increase that doesn't register as inflation.

This extends beyond package sizes into quality degradation that's harder to measure. Restaurants cutting portion sizes, hotels reducing housekeeping, airlines cramming more seats: the price stays the same but you're getting less. The BLS makes "quality adjustments" but they mostly flow one direction—assuming products improve over time even when the opposite is obviously true.

Software subscriptions are the ultimate shrinkflation vehicle. Slack didn't raise prices—they just moved message history limits from unlimited to 90 days on the free tier, then to 10,000 messages, then introduced that limit on paid tiers too. Each degradation pushes users toward higher-priced plans without technically increasing the cost of existing plans.

The Fed Is Fighting Inflation It Can't See

Monetary policy operates on CPI data that increasingly diverges from lived experience. The Fed raised rates to combat inflation that already looked tame in official statistics, which meant overtightening on paper while undertightening in reality. Rate hikes crushed demand for goods that weren't inflating anymore while barely touching services inflation that never stopped.

The disconnect creates policy mistakes in both directions. When the CPI says inflation is controlled, pressure builds to cut rates even though housing, healthcare, and education costs are still climbing. When the CPI spikes on energy prices, the Fed tightens into what might be a temporary shock.

We're stuck with a measurement system designed for an economy where people bought physical goods and inflation meant factories raising prices. That world is gone. The new economy runs on subscriptions, services, and switching costs—and we're still pretending the old metrics tell us anything useful.

What Actually Matters

The real inflation rate is whatever's happening to the things you can't avoid buying. For most people under 40, that's rent, healthcare, childcare, and the software subscriptions their job requires. None of those are captured accurately by the CPI.

The metric isn't wrong because of conspiracy or incompetence—it's wrong because the economy changed faster than our ability to measure it. The BLS is trying to track 2024 spending patterns with a methodology built for 1980. The gap between measurement and reality keeps widening, and we keep making policy decisions based on numbers that describe an economy nobody lives in.

Until we fix the measurement problem, we're flying blind. The inflation rate is whatever the CPI says it is, and the CPI says everything is fine. Meanwhile, everyone under 40 is watching their rent, subscriptions, and childcare costs eat 70% of their income and wondering why the experts keep insisting inflation is under control.