Monthly

Data is collected throughout the month by  Insee collectors. Reports are monthly.

The French oversea departments and Regions (DROM) have been included in the CPI since the 1998 base.

Mayotte has been included in the CPI since 2023.

Three types of data collection are used :

• Part of the collection is carried out by Insee collectors in stores ; tablet computers are used for data input.

• A number of prices are collected from e-commerce websites. This data collection is performed either manually or via bots (harvesting, also known as web scraping).

• Insee finally collects daily scanner data of retail chains; their transmission to Insee was made compulsory by an implemented order of April 13, 2017 making the transmission of data by electronic means compulsory for public statistics purposes. These data are used for the manufactured food products and cleaning and personal care products sold in supermarkets and hypermarkets.

• Other administrative and private data collected as part of the Consumer Price Index survey are leveraged for the calculation of the Consumer Price Index;

• The Rent and Charges survey and the Social Landlord Rent survey are used to measure the rent price index.

There is no sampling plan tracing all household consumption by item, retail outlets, etc.

Voici une proposition de traduction qui respecte le ton institutionnel et la précision statistique :

In the absence of a sampling frame, the individual consumption categories are first broken down by product type. A sample of over 5,000 product families (1,900 for metropolitan France and 300 to 800 per DROM where they are specific), referred to as 'varieties', is thus defined. This allows for the heterogeneity of goods and services within the various consumption items, as defined by the COICOP nomenclature, to be taken into account.

The weightings are based on national accounts data and, at more granular levels, on the Household Budget Survey and other sources.

The sample is updated annually to reflect changing consumption patterns, specifically to introduce new products or, conversely, to stop tracking prices for products whose consumption is declining or even disappearing.

• For field surveys conducted in brick-and-mortar outlets (a sample of approximately 140,000 prices each month), the sample is derived from a two-stage sampling design: (i) in addition to each DROM, 87 urban areas with over 2,000 inhabitants are randomly selected in metropolitan France to ensure representativeness in terms of size and geographical location; (ii) the products to be priced and the outlets to be surveyed are selected using a quota method, requiring enumerators to conduct a specific number of price collections per sales format of the store where the product is observed.

In total, about 30,000 outlets are surveyed each month.

    • For web-based data collection, ad hoc information can sometimes be leveraged regarding the market shares of different operators or the most consumed products (e.g., statistics or data collected on trips for transport services, or on consumer profiles for telecommunications, insurance, and energy contracts). In the absence of such information, the choice of products is made by the survey officer and the price readings are equally weighted.

    • For products monitored using transaction data or scanner data, a sampling frame is available since the precise quantities consumed are known by item and outlet. The exhaustiveness of the data is utilized.