Introduction
TAXIPP is the IPP’s Micro-simulation model for household taxation. It draws on the extensive experience in Micro-simulation techniques at the Paris School of Economics and is maintained by a permanent team at the IPP. TAXIPP is used both in academic research projects and in applied studies that directly inform public debate.
The TAXIPP model aims to simulate, on a representative sample of the French population, the entire monetary redistribution relating to households – that is to say, compulsory levies on the one hand, and social benefits on the other. This model makes it possible to simulate, for each individual in this database, the full range of their tax and benefit payments, both under the current system and under any reformed system. It thus makes it possible to assess the impact of socio-fiscal reforms on both public finances and the distribution of income after redistribution.
Online resources
The TAXIPP model’s
code is open-source. You can view:
The online code: https://gitlab.com/IPP/partage-public-IPP/TAXIPP
Detailed documentation for the model: https://ipp.gitlab.io/partage-public-IPP/TAXIPP/
The validation note:https://www.ipp.eu/en/publication/validation-report-for-the-taxipp-microsimulation-model-version-2-4-2/
and the underlying data
Model architecture
The TAXIPP model has two distinctive features compared with existing models in France. Firstly, it is based on the integration of various databases, most of which are of administrative origin and are comprehensive. Secondly, this model links the database created through this integration to an open-source, collaborative socio-fiscal simulator – the OpenFisca simulator – to which the IPP contributes, and which is intended for shared use across a range of applications (data processing, web APIs, etc.).
The TAXIPP model is essentially based on four data sources: the Fideli files (INSEE), derived from data on residential/local housing tax, income tax and property tax; the Felin files (DGFiP), which provide detailed information from the income tax returns of 500,000 French tax households, offering a comprehensive picture of the highest earners; the DADS files, which contain all the information from employees’ social security declarations for the collection of social security contributions; and the BNS files (INSEE), which are similar to the DADS files but cover the self-employed. Although the TAXIPP code is open source, the data used is protected by statistical confidentiality; access requires the approval of the Statistical Confidentiality Committee and is facilitated via the Centre for Secure Data Access (CASD). TAXIPP uses Python for its simulations and Stata for the process of integrating the various data sources.
Team
Authors of the model (in alphabetical order): Mahdi Ben Jelloul, Paul Dutronc-Postel, Sylvain Duchesne, Brice Fabre, Arthur Guillouzouic–Le Corff, Lola Josseran, Chloé Lallemand, Claire Leroy, Nolwenn Loisel, Fabian Reutzel, Lukas Puschnig.
Current team: Sylvain Duchesne, Paul Dutronc-Postel, Bertrand Garbinti, Lola Josseran
Project manager: Bertrand Garbinti
Support
Taxipp receives support from the CNRS as part of the SOSI projects, as well as funding from the ANR under the ‘University Research Schools’ scheme (ANR-17-EURE-0001).
