ERFS : Tax and social incomes survey - 2023
Data description
The Tax and Social Incomes Survey (ERFS) measures standards of living by combining data from the fourth quarter Employment Survey each year with income from administrative sources.
Sector coverage
ERFS enables statistics to be produced and studies to be carried out on the income, standards of living, poverty, and monetary inequalities of individuals living in « ordinary » households in metropolitan France.
Households where the reference person is a student are excluded from the scope of dissemination.
Statistical concepts and definitions
• Declared income : this is the income reported on form no. 2042 ;
• Initial income : this is all income from employment, replacement income and assets before any taxation ;
• Disposable income : this is the income available to the household for consumption and savings.
The components are explained in the document "Concepts of Tax and social incomes survey / ERFS".
Statistical unit
Individuals and households
Statistical population
"Ordinary" households whose income declared to the fiscal administration is positive or zero and where the reference person is not a student.
Households living in collective dwellings (hostels, prisons, hospitals, etc.) are excluded, as people living in mobile accommodations (e.g. boatmen) and the homeless
Reference area
France métropolitaine
Year N
Comparability - geographical
Not applicable
Comparability - over time and CC2. Length of comparable time series for U
Data has been available annually since 1996.
Source data
- Données issues de l’Enquête emploi en continu (Insee)
- Fichiers sur les déclarations de revenus fiscaux,(Direction générale des finances publiques)
- Fichiers sur les prestations sociales versées par les principaux organismes gestionnaires de ces prestations (Cnaf, Cnav et CCMSA)
Data compilation
See section S.10.6 - Documentation on methodology
Sample size
About 45,000 households
Data collection documents
Not applicable
The survey sample is made up of respondents to the Labour Force Survey (EEC) of the 4th quarter of the year N. A statistical match is made with the tax sources for the year N, i.e. the tax returns for year N (completed in March N+1) and the local residence tax (on 1st January N+1) provided to the INSEE by the fiscal administration.
Some further improvements have been applied to the former Tax Income Survey (ERF) in line with the recommendations of the 2007 report on standards of living and inequalities by the French National Statistical Information Council (CNIS) - which is the body that prepares the annual agenda for the INSEE.
In the Tax Income Survey (ERF), non-taxable social incomes (family-related allowances, housing allowances and minimum social benefits) were deemed to be estimated according to a computation schedule taking into account the type of household or by econometric simulation.
In the new series, called the Tax and Social Incomes Survey (ERFS), the amounts actually received by the households over the reference year N are collected directly from the three main Benefit funds in France : the national family allowances fund (CNAF), the national old-age insurance fund (CNAV) and the central agricultural social insurance fund (CCMSA). These data are matched with the survey - household by household - thanks to a statistical operation of merging.
Moreover, the old series of Tax Income Surveys which relied exclusively on fiscal sources did not properly assess investment incomes, mainly because these incomes are not to be recorded on tax returns, and therefore minimised the income inequalities that could be measured.
In addition to the first improvement, the incomes that are generated by several financial products and are not to be recorded on the tax returns are now imputed using stochastic imputation models.
Imputation models rely on models of financial asset holdings patterns that are first estimated on microdata from the Household Wealth Survey. Incomes from such assets are then computed based on average return rates that are applied to these assets amounts.
In 2021, the Labour Force Survey (EEC) underwent a major overhaul which required a review of the ERFS production chain. Adapting production to this new EEC also provided an opportunity to streamline the entire ERFS production chain and review the method used to correct for non-response.
Some ERFS surveys are supplementedy matching with other administrative sources, for research and expert analysis purposes. Thus, the ERFS 2020 and subsequent surveys are matched with pseudonymised individual data from the Monthly Resource System (DRM), an information system provided for by Decree No. 2025-1288, in the form of intermediate production files (FPI), prototypes whose use is restricted (list of INSEE, CNAF, and DREES members). This matching is carried out by the CNAV (as the DRM operator) on behalf of INSEE. This matching includes monthly information over three years on individuals' income, paying agencies and underlying relationships (employment contract, entitlement to unemployment benefits, etc.) (see favourable opinion issued by the Statistical Confidentiality Committee on 3 October 2023 and decision of the Archives de France on 30 October 2023 – File M1798).