Inflation and CPI: How the Most-Watched Number in Macro Is Actually Made
5 steps · one page
In short
Pillar 1's inflation article covered the concept — rising prices, shrinking purchasing power. This article covers the measurement: the Consumer Price Index, the statistical machine that turns millions of individual prices into the single number that moves markets on release morning.
Measurement matters because the index is not the phenomenon: two people can experience different inflation than the official rate and both be right about their own baskets, and markets react to specific technical cuts of the data — core, month-over-month, versus expectations — that headline reporting flattens. Knowing how the sausage is made turns CPI day from noise into information.
How a price index is built
A statistical agency — the BLS in the US, Eurostat and national offices for the EU's HICP — defines a basket: the goods and services an average household buys, weighted by spending shares (housing, food, transport, services, and so on). Field and digital collection then tracks the prices of those items month after month, and the index is the weighted average of their changes. Everything reported follows: year-over-year inflation compares this month's index with the same month last year (smooth, but slow to show turning points — a big month from a year ago dropping out of the window moves the annual rate with no new inflation at all, the "base effect" behind many confusing headlines); month-over-month shows the fresh impulse (noisy, but current, and the cut markets increasingly parse first). The construction also explains the honest limits, which the agencies themselves document: substitution (households switch away from what got expensive; methods adjust with a lag), quality change (this year's phone is not last year's phone — hedonic adjustments try to price the difference, a methodological choice with critics on both sides), and housing (owner-occupied housing costs are estimated differently across countries — a major reason US and EU inflation aren't perfectly comparable). These are documented methodological trade-offs in a genuinely hard measurement problem — worth knowing, without the conspiracy framing they sometimes attract.
The cuts that markets actually watch
Headline vs core: headline is the whole basket; core strips food and energy — not because they don't matter to households (they do), but because they are volatile and often driven by global supply factors monetary policy can't touch, so core is read as the better signal of underlying, persistent inflation. Central banks say so explicitly, which is why a hot headline with a cool core can produce a counterintuitive market reaction. Which index: the US has two families — CPI (published first, wired into inflation-linked bonds and many contracts) and PCE (broader weights, methodological differences, and the Federal Reserve's stated preferred gauge) — so "inflation" in a US headline can mean two numbers that disagree; the euro area's HICP is the ECB's target measure with national variants beneath it. Versus expectations: as everywhere in this pillar, markets price forecasts in advance and move on the surprise — a 3.4% print against 3.6% expected can rally markets that a 2.9% print against 2.7% expected would sink. The release-day mechanics get their own article in this pillar; the takeaway here is that "CPI came in high" is meaningless without the expected number beside it.
What inflation does to asset classes — structurally
Descriptive, not predictive. Cash and fixed-coupon bonds are inflation's most direct casualties: fixed payments buy less, and the real-vs-nominal arithmetic can turn positive yields into negative real returns — the historical reason inflation-linked bonds exist. Equities are mixed by mechanism: companies with pricing power can pass costs through (revenues inflate alongside expenses), but inflation typically brings the rate rises of the previous article, compressing valuations — so the equity record across inflationary episodes is genuinely uneven, moderate inflation historically coexisting with fine returns and high inflation with difficult ones. Real assets (property, commodities) carry the intuitive inflation-hedge story with a messier empirical record than the story suggests. The literacy point is the pattern: each asset's inflation exposure follows from whether its cash flows are fixed or adjustable in nominal terms — a structural question you can ask of anything you hold.
Worked example
Worked example (fictional). Release morning: consensus expects 3.6% year-over-year, 0.3% month-over-month core. The print: 3.7% headline — but 0.2% core MoM. Headlines say "inflation rises"; markets rally. Nothing is broken: the headline beat came from a volatile energy component and a base effect, while the core monthly impulse — the cut read as the persistent signal — came in below expectations. Two numbers, two audiences: the headline described the household's month; the core surprise described the trend markets price. Reading both is the whole skill this article teaches. All figures are illustrative.
Frequently asked
5 questions
How is CPI actually calculated?
A statistical agency tracks the prices of a weighted basket representing average household spending, and the index is the weighted average of those price changes. Reported inflation is the index's change — year-over-year for the trend, month-over-month for the fresh impulse.
What is core inflation and why exclude food and energy?
Core strips the basket's most volatile, globally-driven components to reveal underlying persistent inflation — the signal central banks say they steer by. It is not a claim that food and energy don't matter; it is a filter for trend versus noise, and markets often react more to core than headline.
Why do CPI and PCE disagree?
Different baskets, weights, and formulas: US CPI is published first and wired into contracts and inflation-linked bonds, while PCE — the Federal Reserve's preferred gauge — weights spending differently and typically runs slightly lower. Both are legitimate; "US inflation" simply has two official answers.
Why did markets rally when inflation came in high?
Almost always: versus expectations, or the core cut. Markets pre-price the consensus forecast, so reactions follow the surprise, not the level — and a hot headline with a soft core reads as noise over signal. Any release reaction is unreadable without the expected number next to the print.
Is official CPI understating my inflation?
Possibly yours, genuinely: the index measures an average basket, and a household whose spending skews toward fast-inflating categories experiences more than the official rate. The documented methodological debates (substitution, hedonics, housing) are real trade-offs statisticians publish openly — a different thing from the index being rigged.
References
- US Bureau of Labor Statistics — CPI Questions and Answers —
- Eurostat — Harmonised Index of Consumer Prices (HICP) —
- US BEA — PCE Price Index —
Educational and informational only — not investment advice, a recommendation, or an offer to buy or sell any security. Investing involves risk, including the possible loss of principal. Worked examples use fictional companies and figures.