Filosofi data are derived from the reconciliation of tax data supplied to Insee by the Direction Générale des Finances Publiques (DGFiP) and data on social benefits from the main organizations managing these benefits: the Caisse Nationale des Allocations Familiales (CNAF), the Caisse Nationale Assurance Vieillesse (CNAV) and the Caisse Centrale de la Mutualité Sociale Agricole (CCMSA).

Tax data (DGFiP):

• The tax returns file (Permanent des Occurrences de Traitement des Émissions, POTE) contains data relating to the year's tax returns (forms 2042 and 2042C) sent by taxpayers to the DGFiP in the spring of the following year. This file lists tax households. Observations corresponding to individual tax returns are extracted.

• The housing tax file (Permanent Local Foncier Commun, PLFC) lists taxpayers connected to premises subject to housing tax (TH) on January 1 of the year following the income year. Observations corresponding to TH taxpayers connected to dwellings taxed as principal or secondary residence are extracted (some households declare income in the département of their secondary residence). These dwellings enable us to move from the notion of tax household to the notion of tax household.

• The Fichier d'Imposition des Personnes, FIP, contains the personal data (surname, first name, date and place of birth, address, etc.) of each taxpayer (known in the sense of POTE and PLFC). This data is used for matching with social security files (once the files have been matched, personal information is not retained). The file is also used to geolocate addresses in order to produce local, sub-municipal statistics.

Social data (Cnaf-Cnav-CCMSA)

• INSEE collects, via CCMSA, the annual amounts of legal benefits for the old-age and family branches of the agricultural scheme.

• For the general scheme, Cnaf provides an exhaustive file of recipients and the amount of their social benefits during the year (FAR6), and Cnav provides an exhaustive file of benefits paid in December. Annual benefit amounts are then reconstituted by extrapolation, using in particular available information on family composition.

Annual

Within the framework of agreements, the DGFIP, Cnaf, Cnav and CCMSA transmit data to Insee.

Data checks are carried out throughout the production process: comparisons with data from the previous year and other sources (ERFS, etc.), quality control of georeferencing data, consistency checks.

Producing Filosofi results requires two pairings:

• matching of the POTE and PLFC tax files, to delimit the scope of Filosofi and create tax households;

• matching between tax sources, and more specifically the FIP file on the one hand, and social sources on the other, to reconcile declared income and taxes with social benefits, in order to arrive at disposable income. Matching is carried out source by source (CNAF, MSAF, MSAV, CNAV).

In Filosofi, as in the old RDL device, non-subject to declaration financial incomes are allocated for metropolitan households according to a pattern built from the Household Wealth Surveys (INSEE): a probability of detention and a detention amount are estimated according to various observable households characteristics (income, age, family situation ...) for seven financial products (tax-exempt savings deposits , LEP, youth passbook savings accounts , CEL, PEL, life-insurance and PEA.