Register
Skip to main content
All you need to know
< All Topics
Created06/04/2026
Updated29/06/2026

ONS Monthly Data on Consumer Prices Index (CPI)

The Monthly Consumer Price Index (CPI) measures the average change in the prices of goods and services purchased by UK households 4 . It is constructed using the internationally recognized Harmonised Index of Consumer Prices (HICP) methodology, ensuring that UK inflation figures can be compared directly with those of European Union nations and other global economies 2, 4 .
  • Methodology and Construction
    • The Basket of Goods and Services: The CPI is based on a representative “shopping basket” of approximately 740 to 760 goods and services 4 . To ensure the index remains representative of contemporary lifestyle trends, the ONS reviews and updates the contents of this basket annually (typically in February) 4 .
    • Price Collection: Every month, the ONS gathers roughly 180,000 price quotes across the UK 4 . These are collected by visiting retail outlets in various locations, scraping online prices, and—increasingly—using point-of-sale “scanner data” provided directly by major retailers.
    • Weights: Not all items in the basket affect the index equally. Items are weighted based on their share of total household expenditure, derived primarily from the ONS Living Costs and Food Survey alongside National Accounts data.
  • Formula and Calculation
    At the lowest level of aggregation (where individual price quotes are combined before weights are applied), CPI primarily uses a geometric mean (specifically the Jevons index formula) 1, 5 
Geometric Mean = x1 × x2 × ⋯ × xn n
Using a geometric mean is economically robust because it implicitly assumes that consumers will substitute goods to some extent when relative prices change (e.g., if the price of one brand of coffee rises dramatically, some consumers will switch to a cheaper competitor) 5 . When economists, analysts, and policymakers look to “deflate” historical economic data (such as adjusting wages, corporate revenues, or GDP to “real” terms), they must choose an appropriate deflator 1 . The choice between CPI and RPI has significant statistical and financial implications.
  • Pros of Using CPI (instead of RPI)
    • No “Formula Effect” Bias: RPI relies heavily on the arithmetic mean (specifically the Carli formula) at its lowest level 4 5 . Mathematically, the arithmetic mean suffers from an upward bias known as the “formula effect.” This bias causes RPI to overestimate inflation by approximately 0.8 to 1.0 percentage point per year compared to CPI, purely due to mathematical aggregation rather than actual price increases 1, 5 . CPI’s geometric mean eliminates this distortion 1, 5 .
    • Mitigation of Substitution Bias: CPI models realistic consumer behavior by assuming that people substitute goods as relative prices shift 5 . RPI assumes zero substitution (representing a rigid basket), which leads to an overestimation of the actual cost of living increases felt by households 1, 5 .
    • International Comparability: CPI conforms to HICP standards 2, 4 . If you correct economic data using CPI, your “real-term” figures are directly comparable to international datasets (e.g., comparing real wage growth in the UK vs. Germany) 2 . RPI is highly idiosyncratic to the UK 2 .
    • Broader Population Coverage: CPI represents a wider population base, including institutional households (such as student accommodation and care homes) and the top 4% of high-income earners 4 . RPI explicitly excludes high-income households and pensioner households that derive the majority of their income from state benefits 4 .
  • Cons of Using CPI (instead of RPI)
    • Exclusion of Housing Costs: The primary disadvantage of CPI is that it completely excludes owner-occupiers’ housing costs (OOH), such as mortgage interest payments, house depreciation, and Council Tax 1, 4 . RPI includes these expenses 1, 4 . If mortgage interest rates rise sharply, CPI will not capture this direct squeeze on household finances, whereas RPI will 4, 5 . (Note: While CPIH addresses this issue, CPI remains the global comparative standard 3 ).
    • Disruption to Historical Continuity: RPI has been calculated continuously since 1947 2 . Standard CPI data only goes back to 1996 (with modeled estimates extending to 1988 or 1950) 3 . For very long-run historical corrections (e.g., comparing economic indicators across the 20th century), RPI provides a more consistent, unbroken native dataset 2 .
    • Legal and Contractual Incompatibility: RPI is hardcoded into many legacy private-sector pension schemes, commercial property leases, and index-linked gilts 1, 2 . Correcting or deflating these specific financial instruments using CPI is often legally impossible without contract renegotiation, and doing so would result in a transfer of wealth away from the recipient (since CPI is structurally lower than RPI) 1 .

Referenced internet pages dealing with ONS (Office of National Statistics) on-line information:
Table of Contents
Getting the Risk Score

The final risk score will be computed from the 4 last Annual Financial Accounts

Close