The Consumer Price Index, or CPI, is a broad measure of the prices households pay for a typical basket of goods and services. It shows how the overall cost of living is changing over time and is commonly used as shorthand for inflation.
Markets quote CPI as percentage changes: month on month to show the latest pace, and year on year to compare prices with the same month a year earlier. Many countries also publish seasonally adjusted figures to smooth out regular patterns such as holiday sales or fuel demand.
What CPI actually measures and what it leaves out
CPI tracks the average price level faced by consumers. The basket usually spans food, energy, housing costs, clothing, transport, health, education, communications, recreation and more. Each category has a weight that reflects its share of household spending, so a rise in rent has more impact than a rise in cinema tickets.
What is inside the index is carefully defined. Price collectors record what consumers actually pay, taking account of promotions and taxes on purchases. How owner-occupied housing is treated varies by country. Some versions include explicit housing costs such as rents and repairs, others impute a rent that homeowners would theoretically pay for their homes, and some add measures of housing costs like insurance and council tax. These design choices can make CPI readings differ across jurisdictions even when underlying price moves look similar.
CPI focuses on consumer-facing items. It does not include asset prices such as shares or house prices, and it does not directly measure the cost of borrowing. Those matter for households, but they sit outside the scope of a consumption price index.
Headline vs core CPI, and other useful cuts
Headline CPI is the full index. Core CPI strips out volatile items such as food and energy to reveal underlying trends that policy makers watch for persistence. Some releases break CPI into goods versus services, or tradables versus non-tradables, which helps analysts see where pressure is building.
Other variants exist. Some countries publish CPI including housing measures or owner-occupiers’ housing costs. There are also trimmed mean or median CPI measures that downweight or remove extreme price moves to reduce noise. The choice of measure depends on what question you are trying to answer: short-term heat, broad trend, or a specific sector’s inflation.
How CPI is built: basket, weights and index maths
Statistical agencies design the basket using surveys of household spending. Weights are refreshed periodically so the index reflects how people actually shop. Prices are gathered every month from stores and online outlets, often for thousands of specific items with detailed product descriptions.
CPI is typically a Laspeyres-type index. In plain language, it asks: what would it cost today to buy last period’s basket? That basket is updated from time to time, and the index is chained so the time series stays continuous. When products change quality or new models arrive, statisticians adjust prices so that the index reflects pure price change, not better features. Methods vary by country, but the goal is the same.
A simple example helps. Suppose the basket has only fuel and groceries, with weights 30% and 70%. If fuel prices rise 10% and groceries rise 2% this month, the CPI move is roughly 0.3 × 10% plus 0.7 × 2% = 4.4% for that period. Real indices use many more categories and techniques to handle substitutions and product changes, but the weighted-average idea is the core.
How traders and investors use CPI releases
CPI days are closely watched because inflation shapes interest-rate expectations. A higher-than-forecast reading can push traders to price faster or larger moves in the central bank’s base rate. That tends to lift government bond yields, often quoted in basis points, and can strengthen the currency if markets expect tighter policy. A softer print usually has the opposite effect.
Bond prices, especially at the front and belly of the curve, can swing on the surprise component of CPI. Inflation-protected securities react to both realised inflation and changes in expected inflation. Equity moves depend on which parts of CPI are hot or cooling. For example, sticky services inflation can worry markets about margins and wages, while easing goods prices might lift consumer discretionary names. None of these reactions is guaranteed, but CPI is one of the most market-moving data points on the calendar.
Beyond the release-day jolt, CPI influences long-run return assumptions. Nominal yields on bonds, discount rates in valuation models and wage negotiations all take a steer from inflation trends.
Reading a CPI release: YoY, MoM, base effects and seasonal adjustment
Year on year (YoY) compares prices with the same month last year. It is the headline number most people quote because it is intuitive and smooths short-term wiggles. The catch is base effects. If prices jumped a year ago, today’s YoY figure can look low even if monthly inflation is firm, and vice versa.
Month on month (MoM) shows the current momentum. It can be volatile, which is why many agencies also publish a seasonally adjusted MoM change to filter regular patterns. Analysts often annualise the MoM rate to get a sense of yearly pace if that month’s trend persisted. The rough method is to compound the monthly change over 12 months rather than multiplying by 12, because compounding reflects how prices build on prior increases.
Releases usually come with detailed tables. Look for contributions by category. A big bump in petrol may add several tenths to the headline even if most other components are quiet, while a broad-based rise across services often signals stickier inflation.
CPI vs PPI, RPI and PCE: what is different
PPI is the Producer Price Index, which tracks prices received by businesses for the goods and services they sell. It captures pipeline pressures that may later pass through to consumers, but it is not a consumer cost measure.
RPI is an older retail price measure used in some places for specific contracts and indexation. It differs in coverage, formula and treatment of housing, and can run higher or lower than CPI depending on period. Many analysts prefer CPI-type measures for policy analysis, but practices vary by jurisdiction.
PCE inflation, common in some countries, uses a broader view of consumption that can reweight spending more frequently and include items purchased on behalf of households, such as healthcare paid by insurers. It often runs differently from CPI because of those methodological choices.
The key point is that these indices are not interchangeable. When comparing across countries or time, check which measure is cited and how it is built.
Limits and common pitfalls
CPI is an average. No household is truly average, so your personal inflation may be higher or lower depending on your mix of spending. It may also lag real-world behaviour if consumers switch to cheaper substitutes faster than the basket update cycle captures.
Quality change and new products are hard to measure. Adjustments aim to strip out the value of better features so the index reflects pure price moves, but there is judgement involved. Housing is another flashpoint. Whether the index uses rents, imputed rents or other housing costs can shift the level and the path of inflation across countries.
Finally, revisions happen. Seasonal factors are updated and weights roll forward, which can slightly change past readings. For trading and risk management, what matters most is the surprise versus expectations, the breadth of the move across categories and how it feeds through to rate path assumptions.
Used with those caveats in mind, CPI remains the central yardstick for consumer inflation and a fixture on every market diary.