How to read a CPI report
Every month, the BLS, ONS, and Eurostat release inflation data. Here is what the numbers mean, what to look at first, and what the media usually gets wrong.
A Consumer Price Index release lands in your feed and the headlines immediately declare inflation is "up" or "down". But one number — the headline rate — tells only part of the story. A release contains dozens of sub-indices covering energy, food, shelter, services, and more. Understanding how to navigate the full report separates a useful data read from a misleading one.
This guide uses the US Bureau of Labor Statistics CPI-U release as the primary example, with notes on the UK ONS CPI bulletin and Eurostat HICP flash where they differ.
Find the 12-month headline rate
The first number in every CPI press release is the 12-month rate: how much the all-items price index changed compared with the same month a year earlier. For the US, this is the CPI-U all-items figure. For the UK it is CPI (not CPIH). For the euro area it is HICP.
The BLS CPI news release puts this number in the first sentence. The ONS CPI bulletin leads with it in the "main points" summary box. For Eurostat, the flash estimate press release puts headline HICP at the top.
Note whether the rate has risen or fallen relative to the previous month — but do not stop there. A rise in headline tells you less than you think until you see what drove it.
Check core CPI next
Core inflation — all items less food and energy — is almost always the second number in any reputable analysis. Food and energy prices are volatile: a single OPEC decision or drought can swing them by several percentage points in a month. Core tells you whether the underlying price pressure in the economy — wages, services, rents — is rising or falling.
In a BLS release, core appears in Table 1 as "All items less food and energy". In the ONS bulletin, it is reported in the same main-points summary. The ECB's Economic Bulletin publishes a detailed breakdown of HICP core components after each Governing Council meeting.
Rule of thumb: if headline and core are moving together, inflation is broad-based. If headline is moving but core is not, it is usually an energy story.
Identify the largest contributor
CPI releases include a contribution table showing how many percentage points each component added to (or subtracted from) the overall rate. This is the most important analytical tool in the release and the one most often ignored by news headlines.
In the BLS release, Table 2 shows the "contribution to change" for each major category. A component running at 10% but with only 3% basket weight contributes less than a component at 4% with 35% weight (shelter). For the US, shelter alone typically contributes 1.5–2.0 percentage points to the total CPI rate — more than any other single category. The BLS CPI detailed tables publish this data for every release.
Look at the monthly change too
The 12-month rate compares today with a year ago. But the monthly change (month-on-month, or MoM) shows the most recent momentum — what happened in the last 30 days. Annualised, the monthly change gives a sense of where inflation is heading.
A 12-month rate of 3.2% with a monthly change of +0.6% (annualised: ~7%) signals acceleration. The same 3.2% rate with +0.1% monthly (annualised: ~1.2%) signals rapid deceleration. Many analysts and the Federal Reserve watch 3-month annualised rates as the most responsive current read. The FRED database makes it easy to calculate these for any time period.
Put it in context: base effects and revisions
A year-on-year rate is always compared against what happened 12 months ago. If inflation was very high last October, this October's rate will look lower by comparison — even if prices have not changed at all. This is the base effect. It is mechanical and expected; good analysis accounts for it.
Also note that initial CPI figures are sometimes revised. The IMF and World Bank both publish revision notes alongside their databases. For the US, BLS CPI is rarely revised after publication — unlike GDP. For the euro area, Eurostat's flash estimate may be revised ±0.1 pts in the final release two weeks later.